5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

8/19/19 2:25 PM

OFFLINE: CHALLENGING INTERNET AND SOCIAL
MEDIA BANS FOR INDIVIDUALS ON SUPERVISION
FOR SEX OFFENSES
JACOB HUTT¥
ABSTRACT
Tens of thousands of people across the United States are subject to bans on
their Internet and social media access due to sex offense convictions. This Article
explains why, even for those on parole and probation, such bans are frequently
overbroad, imposed on the wrong people, and are now ripe for challenge in light
of the Supreme Court’s 8-0 decision in Packingham v. North Carolina. The first
flaw with these bans is their mismatch between crime and condition. They are
imposed on individuals whose criminal records have no relation to online predatory activity or manipulation of minors. The second flaw is their extreme overbreadth. Rather than merely proscribing speech with minors or access to certain
online forums, they cordon off the Internet itself, ostracizing offenders to an offline
society. While these flaws rendered Internet and social media bans constitutionally problematic before the Packingham decision, the Supreme Court’s imprimatur on free speech for individuals convicted of sex offenses could—and should—
lead the way to future legal challenges of these bans.
I. WHAT IS SUPERVISION AND HOW DOES IT RESTRICT SOCIAL MEDIA ACCESS?
................................................................................................................ 668
II. THE WRONG PEOPLE AND THE WRONG SPEECH: UNCONSTITUTIONAL
MISMATCHES IN INTERNET AND SOCIAL MEDIA BANS.......................... 673
A. The Analytical Standard—Narrow Tailoring, By One Name or Another 674
B. Conditions on the Wrong People .............................................................. 677
C. Conditions on the Wrong Speech .............................................................. 682

¥
Law Clerk, Hon. Ronnie Abrams, United States District Court for the Southern District of
New York. I would like to thank Hannah Bloch-Webha, Scott Skinner-Thompson, Esha Bhandari,
Lee Rowland, Astha Sharma Pokharel, Eric Janus, Zawadi Baharanyi, and Ira Ellman for their invaluable feedback on this Article. Thanks as well to Peter Eliasberg and Jennifer Granick, who provided me with my first opportunity to litigate these issues. I would also like to thank the various
RLSC editors who worked on this article: Charlotte Heyrman, Vanessa Vallecillo, Alexia Ramirez,
Maggie Lamb, Efosa Akenzua, Sofia Fernandez Gold, and Jasmine Badreddine. Finally, I am deeply
grateful to the many affected individuals on supervision, named and unnamed in this Article, who
have entrusted me with their stories. All errors are my own.

663

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

664

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

1. The Range of Speech Burdened by Internet and Social Media Restrictions
............................................................................................................... 683
2. The Lack of Internet Alternatives .......................................................... 686
3. Other Means of Furthering the Governmental Interest .......................... 687
III. PACKINGHAM AND ITS BENEFICIARIES ........................................................ 691
A. The Packingham Decision ........................................................................ 691
B. Who Does Packingham Benefit? .............................................................. 692
C. How Does Packingham Strengthen Parole and Probation Challenges? ... 698
1. Packingham Says That Social Media Bans Are a Per Se Impingement on
First Amendment Rights. ...................................................................... 699
2. Packingham Suggests That “Social Media” is Too Broad a Category .. 703
IV. CONCLUSION................................................................................................ 704
“Sherman D Manning ‘Preaching’” is worth a watch.1 It’s a twelve minute,
forty-six second long video of Mr. Sherman Manning, a Baptist preacher, bellowing out the biblical story of Samson as a lesson in resilience. He tells a rapt congregation of how Samson—the weakened, imprisoned warrior who ultimately regained his strength—was transformed through adversity: “They did not know that
the same man they put into the dungeon was not the same man that was coming
out of the dungeon. . . . When you go through a fire and come out of it, you’re not
the same person anymore! . . . If you make it out alive, you’re stronger than you
used to be!”2 Manning speaks from personal experience. Before he became a pastor at the Yes We Can! Worship Center, he himself was incarcerated (wrongfully,
in his view)—a period of “darkness” and tremendous growth.3 Since his release
from prison, he has posted online clips of these sermons, in part, to advertise his
preaching skills and gain invitations from local churches in the Los Angeles area
to deliver guest sermons.4 He has been an ordained minister since the age of 18.5
Yet the act of posting the above video on YouTube could have gotten Manning sent back to prison. His convictions—over two decades old—are for sex offenses, meaning that when he was released on parole, his release conditions included a flat ban on the use of “social media.” More specifically: “You shall not
use or access social media sites, social networking sites, peer-to-peer networks, or

1. Yes We Can! Worship Center, Sherman D. Manning “Preaching,” YOUTUBE (Feb. 6,
2017), https://www.youtube.com/watch?v=vM-96Va7m8Y (on file with author).
2. Id.
3. See generally Memorandum in Support of Plaintiff’s Motion for Preliminary Injunction,
Manning v. Powers, No. 2:17-cv-7832 (C.D. Cal. Oct. 30, 2017) (hereinafter “Manning Memo.”).
Disclosure: The author worked on Mr. Manning’s legal team in challenging his parole condition.
4. See Manning Memo. at 4–5.
5. See id. at 4.

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

665

computer or cellular instant message systems; e.g. Facebook, Instagram, Twitter,
Snapchat, Lync, Gmail, Yahoo, KIK messenger, Tumblr, etc. This would include
any site which allows the user to have the ability to navigate the internet undetected.”6 Manning found this condition puzzling: his crimes had not involved child
pornography, solicitation of a minor online, or any other activity involving use of
the Internet. Yet in order to stay out of prison, he was prohibited from sharing
videos of himself vibrantly engaged in his profession.
Internet and social media bans currently affect tens of thousands of individuals conditionally released from prison on parole, probation, or supervised release—collectively, “supervision”7—who are blocked from using social media as
a condition of being free from incarceration.8 Bans are most frequently imposed
on individuals convicted of sex offenses, whether or not these offenses involved
the Internet or any predatory activity. In some states, judges and supervision officers do not even have the option of declining to impose these access restrictions,
as state law requires social media or Internet bans for all those convicted of sex
offenses who are under state supervision.9 The irony of these restrictions lies in
the supposedly rehabilitative and reintegrative purposes underlying supervision:
the very technology that supervised individuals could use to seek out employment,
to reconnect with estranged family members, to become engaged in politics, and
to stay informed on current events is prohibited. The ends of parole—“to help
individuals reintegrate into society as constructive individuals as soon as they are

6. Manning v. Powers, 281 F. Supp. 3d 953, 957 (C.D. Cal. 2017).
7. Unless otherwise noted, this Article uses the term “supervision” throughout this piece as a
catch-all for parole, probation, and supervised release, in line with other scholarship on this subject.
See generally Cecelia Klingele, Rethinking the Use of Community Supervision, 103 J. Crim. L. &
Criminology 1015 (2013) (exploring diverse iterations of “community supervision”); Cecelia
Klingele, The Role of Sentencing Commissions in the Imposition and Enforcement of Release Conditions, 26 Fed. Sent. R. 191, 191 (2014) (referring to probation and post-release supervision collectively as “community supervision”).
8. This is a conservative estimate. Over ten thousand individuals in New York state alone are
subject to a ban on accessing “commercial social networking websites” due to their sex offenses.
See N.Y. Executive Law § 259-c(15) (banning any “level three” sex offender parolee, as well as any
sex offender parolee whose victim was under 18 or whose crime involved use of the Internet, from
using social media); N.Y. Penal Law § 65.10(4-a)(b) (same for probation); Registered Sex Offenders
by County, N.Y. STATE DIVISION OF CRIMINAL JUSTICE SERVS., http://www.criminaljustice.ny.gov/nsor/stats_by_county.htm [https://perma.cc/C43M-CFXD] (last updated Feb. 5, 2019)
(identifying 10,289 individuals as “level three” sex offenders). Even if one calculated the total number of individuals subject to mandatory bans, this would enormously underestimate the total number
of individuals barred from the Internet, as it would leave out the class of individuals banned via
supervisory officer discretion. See, e.g., J.I. v. New Jersey State Parole Bd., 228 N.J. 204, 210–11
(2017) (invalidating Internet ban imposed by the District Parole Supervisor).
9. See, e.g., 730 ILL. COMP. STAT. ANN. 5/5-6-3.1(t) (“An offender placed on supervision for a
sex offense as defined in the Sex Offender Registration Act committed on or after January 1, 2010
. . . shall refrain from accessing or using a social networking website as defined in Section 17-0.5 of
the Criminal Code of 2012.”).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

666

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

able”—are obstructed by its means.10 The result is a burgeoning class of individuals who seek reintegration into society but are prevented from doing so by their
conditions of supervision.
This Article contends that these Internet and social media bans are often unconstitutional under the First Amendment, and are ripe for challenge after the recent decision in Packingham v. North Carolina.11 Contrary to the general rule that
regulations burdening speech must be narrowly tailored to further a governmental
interest, many of these bans target both the wrong people and the wrong speech.
They are applied to individuals whose criminal records suggest no inclination to
online predatory activity, or to predatory activity at all. And rather than targeting
speech with a clear nexus to criminal recidivism, they sweep broadly in proscribing speech that is at the very least harmless and at most indispensable for reintegrating into society. The Packingham decision, where a unanimous Supreme
Court struck down a North Carolina law making it a felony for any individual
classified as a “sex offender” to access social media, rendered even more vulnerable a wide array of similar restrictions on individuals under supervision, such as
Manning.12
The need for this scholarship is twofold. First, this Article adds to the relatively sparse literature on the constitutionality of restrictions on Internet access as
a term of supervision.13 Other scholarship in this area has focused on conditions
that target those convicted of child pornography crimes14 and conditions that target all “sex offenders,” not just individuals on probation, parole, or supervised
10. Morrissey v. Brewer, 408 U.S. 471, 477 (1972).
11. 137 S. Ct. 1730 (2017).
12. Manning v. Powers, 281 F. Supp. 3d 953, 960 (C.D. Cal. 2017) (“The court finds that this
case is controlled by the Supreme Court’s recent decision in Packingham v. North Carolina.”).
13. See generally Jane Adele Regina, Access Denied: Imposing Statutory Penalties on Sex
Offenders Who Violate Restricted Internet Access as a Condition of Probation, 4 SETON HALL CIR.
REV. 187 (2007) (assessing Fourth Amendment implications of remote monitoring for individuals
under supervision and advocating independent civil and criminal offenses as punishment for violating internet access conditions); Emily Brant, Sentencing ‘Cybersex Offenders’: Individual Offenders
Require Individualized Conditions When Courts Restrict Their Computer Use and Internet Access,
58 CATH. U. L. REV. 779 (2009) (exploring the spectrum of deference given to sentencing courts in
fashioning computer and Internet restrictions as conditions of supervision); Elizabeth Tolon, Updating the Social Network: How Outdated and Unclear State Legislation Violates Sex Offenders’ First
Amendment Rights, 85 FORDHAM L. REV. 1827 (2017) (analyzing legislatively mandated supervision
conditions); Gabriel Gillett, A World Without Internet: A New Framework for Analyzing A Supervised Release Condition That Restricts Computer and Internet Access, 79 FORDHAM L. REV. 217
(2010) (assessing supervision conditions through a novel approach to the unconstitutional conditions
doctrine).
14. See generally Christopher Wiest, The Netsurfing Split: Restrictions Imposed on Internet
and Computer Usage by Those Convicted of a Crime Involving a Computer, 72 U. CIN. L. REV. 847
(2003); Elizabeth P. Evans, Internet Access Restrictions for Convicted Child Pornography Sex Offenders: How Far is too Far?, 36 AM. J. TRIAL ADVOC. 329 (2012); Laura Tatelman, Give Me Internet or Give Me Death: Analyzing the Constitutionality of Internet Restrictions as a Condition of
Supervised Release for Child Pornography Offenders, 20 CARDOZO J.L. & GENDER 431 (2014).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

667

release for sex offenses.15 Second, given the already visible split over whether and
how Packingham applies in the supervision context,16 this Article is the first to
explain why it does. This is the first examination of Packingham in the area of
parole, probation, and supervised release, unlike other early analyses of the opinion which have focused on its treatment of the Internet as a public forum.17
The Article begins in Part I with an overview of supervision in the United
States. It first explains what types of statuses fit into “supervision,” and then describes the various ways supervision conditions are implemented.
Part II describes two legally infirm aspects of supervision conditions that restrict Internet and social media access. At the outset, this requires sifting through
the web of analytical frameworks that can apply to challenges of supervision conditions—among them, strict or intermediate First Amendment scrutiny, modified
constitutional analysis for individuals without full civil liberties, and a unique
standard for supervision conditions that touch on fundamental rights. Rather than
identifying the appropriate standard for each type of challenge, the Article distills
several principles that inform any analysis of supervision conditions.
Part II continues by discussing how supervision conditions restricting Internet
and social media access are legally problematic in two main ways. First, these
conditions frequently target the wrong people. Such restrictions arise both from
laws mandating Internet and social media bans on all “sex offenders” (including
those whose sex offenses were not predatory, were not committed with the Internet, or were not committed against a member of a vulnerable class), and from
discretionary choices by sentencing judges and supervision officers. Second,
many such conditions sweep in the wrong speech.18 They are overbroad for three
15. See generally Jasmine S. Wynton, Myspace, Yourspace, but Not Theirspace: The Constitutionality of Banning Sex Offenders from Social Networking Sites, 60 DUKE L.J. 1859 (2011).
16. See infra Part III.
17. Scholarship examining Packingham has focused generally on its treatment of the Internet
as a public forum, see generally First Amendment-Freedom of Speech-Public Forum Doctrine-Packingham v. North Carolina, 131 HARV. L. REV. 233 (2017), and, more specifically, on the difficulty
the decision could pose to future regulations of misinformation on the Internet, see Richard L. Hasen,
Cheap Speech and What It Has Done (to American Democracy), 16 FIRST AMEND. L. REV. 200, 225
(2017) (raising concern that Packingham will be used “to argue against the constitutionality of laws
that would limit the ability of foreign governments to spread false election-related information to
American voters via social media”) and Eric Emanuelson, Jr., Fake Left, Fake Right: Promoting an
Informed Public in the Era of Alternative Facts, 70 ADMIN. L. REV. 209, 221 (2018) (expressing
skepticism that the government can prevent a disseminator of “fake news” from accessing social
media in light of Packingham).
18. Throughout this Article, I refer to the right to free speech as inclusive of a right to receive
information. See First Nat’l Bank of Bos. v. Bellotti, 435 U.S. 765, 783 (1978) (“[T]he First Amendment goes beyond protection of the press and self-expression of individuals to prohibit government
from limiting the stock of information from which members of the public may draw.”); Citizens
United v. FEC, 558 U.S. 310, 336 (2010) (“[L]aws enacted to control or suppress speech may operate
at different points in the speech process.”); 9 WRITINGS OF JAMES MADISON 103 (G. Hunt ed. 1910)
(“A popular Government, without popular information, or the means of acquiring it, is but a Prologue

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

668

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

main reasons: they sweep in a wide variety of protected speech; they leave insufficient offline alternatives to online speech, and to social media in particular; and
there are several less restrictive means of serving the governmental interest at issue.
Part III explains how Packingham clarifies the constitutional problem with
these bans. First, Part III provides an overview of Packingham v. North Carolina,
and examines the threshold question of whether the decision applies to individuals
under supervision for sex offenses. There are three reasons the answer is yes: first,
the text of the opinion indicates broad application, notwithstanding the Court’s
refusal to rule on whether the fact that the North Carolina law only targeted sex
offender registrants made a constitutional difference; second, permissible restrictions on the civil liberties of sex offender registrants are highly similar to those
imposed on supervisees, therefore these classes should receive similar protections;
and third, the Packingham analysis, with its emphasis on narrow tailoring, fits
neatly into existing doctrine on judicial scrutiny of supervision conditions.
The Article concludes with several ways in which the Packingham decision
supports challenges to broad restrictions on Internet and social media access, even
for those on parole and probation. It offers considerations for why, combined with
nascent case law supporting challenges to these restrictions, this unanimous Supreme Court decision should safeguard the First Amendment rights of parolees
and probationers.
I.
WHAT IS SUPERVISION AND HOW DOES IT RESTRICT SOCIAL MEDIA ACCESS?
Supervision—the catch-all term for probation, a sentence imposed in lieu of
imprisonment, and parole (or supervised release in the federal system19), a term
following a period of imprisonment—is a vast category within the criminal legal
system. Estimates suggest that “more than one-third of those admitted to prison []
arrive there as a result of revocation from community supervision.”20 And this is
to a Farce or a Tragedy; or, perhaps both.”). See generally T.S. Emerson, Legal Foundations of the
Right to Know, 1976 WASH. U. L.Q. 1 (discussing First Amendment right to know).
19. In 1984, amid bipartisan dissatisfaction with federal parole—on the right, disillusionment
with the ability of federal parole to “reform” defendants; on the left, concern that the highly discretionary parole system worked disproportionately against minorities and poor people—Congress
abolished it. See Fiona Doherty, Indeterminate Sentencing Returns: The Invention of Supervised
Release, 88 N.Y.U. L. REV. 958, 991–95 (2013) (detailing the rise of the bipartisan “determinacy
movement” in abolishing federal parole). In its place, Congress established the post-release system
of federal supervised release. See Sentencing Reform Act of 1984, Pub. L. No. 98-473, 98 Stat. 1987
(codified as amended at 18 U.S.C. §§ 3551-3742 (2006)). Supervised release was designed to provide “certainty in release date.” S. REP. NO. 98-225, at 56 (1983), reprinted in 1984 U.S.C.C.A.N.
3182, 3239 (“Under the bill, the sentence imposed by the judge will be the sentence actually
served.”). Supervised release, in practice, “has come to mimic the experience of being on parole or
on probation.” Doherty, supra at 1012.
20. See Klingele, Community Supervision, supra note 5, at 1019.

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

669

a relatively new phenomenon: between 1977 and 2010, the number of individuals
on probation alone grew from 800,000 to more than 4 million.21 Being placed
under supervision means receiving a list of conditions, or the do’s and don’ts for
staying out of prison. These are imposed either by a court or a supervisory body
(usually a parole or probation office), and both types of entities enjoy a high level
of discretion in imposing conditions.22 Across jurisdictions, the relevant authority
typically places a standard set of conditions on a supervisee, and then adds special
conditions based on the individual’s offense, background, and the various goals of
supervision, as discussed below. In the realm of sex offenses, a handful of states
require courts to impose specific conditions, such as restrictions on the individual’s ability to use the Internet or social media, if the supervisee before the court
has been convicted of a sex offense. These mandatory restrictions currently exist
in six states,23 though they vary in terms of who is covered by the law,24 the type
of access restricted,25 and the availability of exceptions.26 More common than
mandatory restrictions are laws providing restrictions that courts “may” impose
on supervisees.27 Less common are quasi-mandatory restrictions, requiring the
court to impose Internet and social media restrictions on anyone under “intensive
supervised release,” leaving it to the court’s discretion whether a person belongs
21. Id. at 1018.
22. See generally Christine S. Scott-Hayward, Shadow Sentencing: The Imposition of Federal
Supervised Release, 18 BERKELEY J. CRIM. L. 180 (2013) (critiquing sentencing judges for failing to
exercise their broad discretion in fashioning conditions of supervised release); James M. Binnall,
Divided We Fall: Parole Supervision Conditions Prohibiting “Inter-Offender” Associations, 22 U.
PA. J.L. & SOC. CHANGE 25, 60 (2019) (50-state parole appendix with various references to discretion
for parole commissions and officers in determining conditions).
23. These states are New York, New Jersey, South Carolina, Texas, Illinois, and Nevada. See
infra note 92 and accompanying text.
24. Compare 730 ILL. COMP. STAT. ANN. 5/5-6-3.1(t) (covering any “offender placed on supervision for a sex offense as defined in the [Illinois] Sex Offender Registration Act committed on
or after January 1, 2010”), with N.Y. PENAL LAW § 65.10(4-a)(b) (covering any individual required
to register as a sex offender where “the victim of such offense was under the age of eighteen at the
time of such offense or such person has been designated a level three sex offender . . . or the internet
was used to facilitate the commission of the crime”).
25. Compare S.C. CODE ANN. § 23-3-555(D) (restricting the access of “social networking
websites” without further definition), with TEX. GOV’T CODE ANN. § 508.1861(b)(2) (restricting the
accessing of “commercial social networking site[s],” defined elsewhere as “an Internet website that:
(A) allows users, through the creation of Internet web pages or profiles or other similar means, to
provide personal information to the public or other users of the Internet website; (B) offers a mechanism for communication with other users of the Internet website; and (C) has the primary purpose
of facilitating online social interactions; and (2) does not include an Internet service provider, unless
the Internet service provider separately operates and directly derives revenue from an Internet website . . . .” TEX. CRIM. PROC. CODE ANN. § 62.0061(f)).
26. Compare NEV. REV. STAT. ANN. § 213.1245(3) (permitting the State Board of Parole Commissioners to exempt a parolee from the Internet ban in “extraordinary circumstances” where “the
Board states those extraordinary circumstances in writing”), with 730 ILL. COMP. STAT. ANN. 5/5-63.1 (allowing no exceptions to the social media ban).
27. See, e.g., FLA. STAT. ANN. § 948.03(1)–(2); N.D. CENT. CODE ANN. § 12.1-32-07(4)(r).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

670

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

in this category.28
Regardless of how Internet and social media restrictions become supervision
conditions, they are an increasingly pervasive way of regulating the activity of
those under supervision.29 Such restrictions have proliferated in the last few years,
often outpacing federal legislative restrictions on “sex offenders.” The federal restrictions were codified in 2006, when Congress enacted Title I of the Adam
Walsh Child Safety and Protection Act of 2006,30 and within it the Sex Offender
Registration and Notification Act,31 or SORNA, which required all U.S. jurisdictions to create sex offender registration systems or risk losing federal funds.32
SORNA’s expansion in 2008 required states to mandate all individuals classified
as “sex offenders” to report their “Internet identifiers,” including email addresses
and other designations used for self-identification on the Internet, to the sex offender registry.33 The statute does limit dispersal of this information: for example,
states are explicitly prohibited from publishing these Internet identifiers on their
publicly available sex offender registry websites.34 Yet in all other respects, much
to the dismay of criminal defense advocates,35 SORNA only establishes “minimum standards,” beyond which states are free to impose more stringent restrictions.36 Beyond the notification requirements of SORNA, state legislatures
28. See, e.g., MINN. STAT. ANN. § 244.05(6).
29. STEPHEN E. VANCE, FED. JUDICIAL CTR., SUPERVISING CYBERCRIME OFFENDERS THROUGH
COMPUTER-RELATED CONDITIONS: A GUIDE FOR JUDGES 1 (Oct. 2015), https://www.fjc.gov/sites/default/files/2015/Supervising%20Cybercrime%20Offenders.pdf
[https://perma.cc/892K-CLD3]
(“Over the past fifteen years, federal district judges have increasingly imposed special conditions of
supervised release and probation restricting computer and Internet use in an effort to protect the
public from cybercrime, including child pornography offenses.”).
30. Pub. L. No. 109-248, 120 Stat. 587 (originally codified at 42 U.S.C. §§ 16901–16991).
31. 42 U.S.C. §§ 16901–16962.
32. Id.; see also Wayne A. Logan, Criminal Justice Federalism and National Sex Offender
Policy, 6 OHIO ST. J. CRIM. L. 51, 82 (2008) (discussing the financial pressure exerted on states by
the Adam Walsh Act).
33. 34 U.S.C.A. § 20916(a) (originally codified at § 16915a) (“The Attorney General, using
the authority provided in section 114(a)(7) of the Sex Offender Registration and Notification Act,
shall require that each sex offender provide to the sex offender registry those Internet identifiers the
sex offender uses or will use of any type that the Attorney General determines to be appropriate
under that Act.”); see, e.g., ARK. CODE ANN. § 12-12-908 (West) (requiring registrant’s file to include “[a]ll social media account information”).
34. See 34 U.S.C.A. § 20916(c); see also OFFICE OF JUSTICE PROGRAMS, U.S. DEP’T OF JUSTICE,
SEX OFFENDER REGISTRATION AND NOTIFICATION ACT SUBSTANTIAL IMPLEMENTATION CHECKLIST—
REVISED
(last
visited
March
30,
2018),
https://ojp.gov/smart/pdfs/checklist.pdf
[https://perma.cc/V9RJ-XHQ2 ] (“The Kids Act of 2008 (42 U.S.C. § 16915a & b.) amended the
SORNA provisions of the Adam Walsh Act by adding Internet identifiers as items that are NOT
permitted to be displayed on sex offender public websites.”).
35. See, e.g., Jacob Frumkin, Perennial Punishment? Why the Sex Offender Registration and
Notification Act Needs Reconsideration, 17 J.L. & POL’Y 313, 348 (2008) (noting the efforts of the
National Juvenile Justice Network to push states not to exceed the federal guidelines in SORNA
regarding juvenile sentencing).
36. See SMART General FAQs, OFFICE OF SEX OFFENDER SENTENCING, MONITORING,

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

671

may impose mandatory restrictions on individuals under supervision,37 and state
supervision departments may standardize special conditions of supervision prohibiting social media access.38 It is exceedingly difficult to know precisely how
many individuals are subject to these social media restrictions, given the decentralized, state-to-state nature of supervision, but mandatory restrictions alone—
that is, without including any cases where courts or supervisory authorities impose
Internet and social media restrictions in their discretion—cover tens of thousands
of individuals.39
Thinking through whether Internet restrictions on parolees and probationers
are fair depends in part on the purpose of parole and probation, and how much
autonomy individuals under supervision are supposed to have. While supervisees,
by definition, are not incarcerated, they are still serving out a criminal sentence.
Yet parolees and probationers retain some civil liberties, with the latter group generally enjoying more than the former. In Samson v. California, the Supreme Court
established that on the “continuum” of state-imposed punishment, individuals under probation enjoy greater civil liberties than those on parole.40 The concept of a
“continuum” of state-imposed punishment means that an individual’s degree of
liberty shifts based on his specific criminal status.41
This “continuum” idea also means that the state’s interests shift depending on
whose liberty is restricted. This is a crucial feature of supervision: unlike incarceration, it is imposed—at least theoretically—to further the individual’s
APPREHENDING, REGISTERING, AND TRACKING (last visited June 6, 2019), https://www.smart.gov/
faq_general.htm [https://perma.cc/9NE8-KU3E] (noting that, beyond the requirement that states exclude certain identifying info, such as Social Security Numbers, from online sex offender registries,
“[i]n all other respects, state discretion to go further than the SORNA minimum is not limited”).
37. See, e.g., N.Y. EXEC. LAW § 259-c(15) (parole); N.Y. PENAL LAW § 65.10 (probation); S.C.
CODE ANN. § 23-3-555 (parole and probation); TEX. GOV’T CODE ANN. § 508.1861 (parole only);
730 ILL. COMP. STAT. ANN. 5/5-6-3.1 (“supervision”); NEV. REV. STAT. ANN. § 213.1245 (parole);
NEV. REV. STAT. ANN. § 176A.410 (probation).
38. See, e.g., Additional Conditions of Supervision for Adult Sex Offenders, COLO. SEX
OFFENDER MGMT. BD. (2018), http://www.advocates4change.org/wp-content/uploads/2018/11/
PROBATION_2018_Additional-Conditions-of-Supervision-for-Adult.pdf [https://perma.cc/C53XG9TA] (The additional conditions include an agreement: “I will not access or utilize, by any means,
any commercial social networking site except under circumstances approved in advance and in writing by the probation officer in consultation with the community supervision team.”) (on file with
author).
39. See supra note 8 and accompanying text.
40. Samson v. California, 547 U.S. 843, 850 (2006) (“On this continuum [of state-imposed
punishments], parolees have fewer expectations of privacy than probationers, because parole is more
akin to imprisonment than probation is to imprisonment.”); id. (“[O]n the Court’s continuum of possible punishments, parole is the stronger medicine; ergo, parolees enjoy even less of the average
citizen’s absolute liberty than do probationers.”) (quoting United States v. Cardona, 903 F.2d 60, 63
(1st Cir. 1990)).
41. For further discussion of this continuum theory of criminal sanction, see NORVAL MORRIS
& MICHAEL TONRY, BETWEEN PRISON AND PROBATION: INTERMEDIATE PUNISHMENTS IN A RATIONAL
SENTENCING SYSTEM, 40–41 (1991).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

672

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

rehabilitation, among other non-retributive goals. In the federal supervised release
context, a court’s imposition of supervised release conditions cannot be justified
by a punitive rationale at all—just deserts, by law, are simply not an aim of supervised release. Instead, pursuant to 18 U.S.C. § 3583, when a federal sentencing
court imposes supervised release conditions, the only interests it may consider are
deterrence, prevention of crime, and rehabilitation.42 The absence of “punishment” or “retribution” from this list of interests is not a statutory oversight.43 As
the Supreme Court has explained, “Congress intended supervised release to assist
individuals in their transition to community life. Supervised release fulfills rehabilitative ends, distinct from those served by incarceration.”44 For this reason, as
the Supreme Court further explained, “‘[s]upervised release . . . is not punishment
in lieu of incarceration,”45 but rather is mainly about “facilitat[ing] the reintegration of the defendant into the community.”46 This may be contrasted with a term
of incarceration, which Congress has explicitly determined should not be supported by a desire to rehabilitate an offender.47 And while probation, at least in
the federal context, may be justified in part by a punitive rationale,48 the Supreme
Court has referred to rehabilitation and deterrence as the “primary goals” of

42. The general sentencing statute, 18 U.S.C § 3553, lists numerous factors which a court shall
consider when it imposes a sentence: the nature and circumstances of the offense and the history and
characteristics of the defendant (§ 3553(a)(1)), promotion of respect for the law and provision of just
punishment for the offense ((a)(2)(A)), deterrence ((a)(2)(B)), prevention of crime ((a)(2)(C)), rehabilitation ((a)(2)(D)), and several others factors. By contrast, when a court is deciding what types of
supervised release conditions it should impose on a defendant, § 3583 says the court may only impose conditions which are “reasonably related to” factors “set forth in section 3553(a)(1), (a)(2)(B),
(a)(2)(C), and (a)(2)(D).” In this list, (a)(2)(A)—permitting the sentence “to reflect the seriousness
of the offense, to promote respect for the law, and to provide just punishment for the offense”—is
conspicuously absent.
43. See United States v. Eaglin, 913 F.3d 88, 99 (2d Cir. 2019) (omitting just punishment from
list of permissible interests in imposing condition of supervised release); United States v. Miller, 634
F.3d 841, 844 (5th Cir. 2011) (“Congress deliberately omitted that factor from the permissible factors
enumerated in the statute.”). As the Fifth Circuit noted, a circuit split has emerged with regard to
reliance on § 3553(a)(2)(A) (the retributive element) under § 3583(e) (the framework for imposing
conditions). Id.
44. United States v. Johnson, 529 U.S. 53, 59 (2000).
45. United States v. Murray, 692 F.3d 273, 280 (3d Cir. 2012) (citing United States v. Granderson, 511 U.S. 39, 50 (1994)).
46. United States v. Vallejo, 69 F.3d 992, 994 (9th Cir.1995).
47. See 28 U.S.C.A. § 994(k) (“The [Sentencing] Commission shall insure that the guidelines
reflect the inappropriateness of imposing a sentence to a term of imprisonment for the purpose of
rehabilitating the defendant or providing the defendant with needed educational or vocational training, medical care, or other correctional treatment.”).
48. Compare U.S. SENTENCING GUIDELINES MANUAL § 5B1.3(b) (2018) (probation condition
can be imposed which is reasonably related to providing just punishment), with id. § 5D1.3(b) (“just
punishment” absent from rationales supporting a supervised release condition). See also United
States v. Brady, No. 02-CR-1043, 2004 WL 86414, at *8–9 (E.D.N.Y. Jan. 20, 2004) (defending use
of a punitive rationale in imposing a term of probation as an alternative to incarceration).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

673

probation.49 Outside the federal context, too, the “traditional view” of probation
has been primarily rehabilitation-oriented, even if a focus on retribution has
emerged in some jurisdictions.50 Without question, restrictions for probationers
can lean, at least in part, on the punitive rationale, whereas restrictions in the nonpunitive supervised release system cannot—but if an Internet ban or similar condition restricts too much of a supervised releasee’s liberty, the same will almost
certainly be true for an individual on probation.51 After all, as noted above, probationers generally enjoy more liberty than parolees.
In sum, supervision is driven largely by non-punitive governmental interests.
As discussed below, this has significant consequences for determining whether a
supervision condition is narrowly tailored to further these interests.
II.
THE WRONG PEOPLE AND THE WRONG SPEECH: UNCONSTITUTIONAL
MISMATCHES IN INTERNET AND SOCIAL MEDIA BANS
Generally, two constitutional flaws exist with supervision’s Internet and social media bans: they target the wrong people and the wrong speech. That is, some
conditions improperly target individuals whose crimes have no relation to their
49. United States v. Knights, 534 U.S. 112, 119 (2001). Specifically, the Court in Knights
listed “rehabilitation and protecting society from future criminal violations” as the “primary goals”
of probation. Id. Although Knights dealt with a condition of probation in California, this reference
to probation’s “primary goals” was not made in construing California penal law, which included
more than just rehabilitation and deterrence among its goals. See CAL. PENAL CODE § 1203.1 (West)
(“The court may impose and require any or all of the above-mentioned terms of imprisonment, fine,
and conditions, and other reasonable conditions, as it may determine are fitting and proper to the end
that justice may be done, that amends may be made to society for the breach of the law, for any
injury done to any person resulting from that breach, and generally and specifically for the reformation and rehabilitation of the probationer.”). Rather, this statement in Knights represents a broader
understanding of the purpose of probation. See Griffin v. Wisconsin, 483 U.S. 868, 880 (1987) (“[I]t
is the very assumption of the institution of probation that the probationer is in need of rehabilitation.”); 18 U.S.C.A. § 3563(b); S. REP. NO. 98-225, at 76 (1983), reprinted in 1984 U.S.C.C.A.N. at
3275 (“Rehabilitation is a particularly important consideration in formulating conditions for persons
placed on probation.”). For a discussion of the interplay between rehabilitative and punitive purposes
in probation, see Wayne A. Logan, The Importance of Purpose in Probation Decision Making, 7
BUFF. CRIM. L. REV. 171, 198–99, 208 (2003).
50. NEIL P. COHEN, THE LAW OF PROBATION AND PAROLE § 1:6 (2d ed. 1999).
51. The insight that not everything housed in the criminal code is punitive is affirmed by the
Ex Post Facto doctrine. The Ex Post Facto clause of Article I, Section 10 of the U.S. Constitution
“forbids the application of any new punitive measure to a crime already consummated.” Lindsey v.
Washington, 301 U.S. 397, 401 (1937). Part of the inquiry in determining whether a measure is
punitive involves asking whether Congress “intended a civil, not a criminal, sanction,” which at first
blush seems to equate “criminal” with “punitive.” United States v. One Assortment of 89 Firearms,
465 U.S. 354, 363 (1984) (quoting Helvering v. Mitchell, 303 U.S. 391, 402 (1938)). See also Smith
v. Doe, 538 U.S. 84, 96 (2003) (incorporating language from 89 Firearms, a Double Jeopardy case,
to apply to Ex Post Facto cases). In Smith v. Doe, however, the Supreme Court explained that “partial
codification of [the Alaska Sex Offender Registration Act] in the State’s criminal procedure code is
not sufficient to support a conclusion that the legislative intent was punitive.” Id. at 95.

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

674

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

offense or background, or to various sentencing goals (deterrence, incapacitation,
rehabilitation, reintegration, etc.); while other conditions, even if targeting the
right people, improperly target speech that need not be proscribed to further the
goals of sentencing. This Part situates these conditions in statutory and constitutional doctrine, explaining how they fail even moderate judicial scrutiny.
A. The Analytical Standard—Narrow Tailoring, By One Name or Another
Before assessing the substantive arguments against an Internet or social media
ban, there is the question of how a court would review such a challenge. Are these
supervision conditions analyzed under standard First Amendment analysis? If so,
do they receive intermediate or strict scrutiny? Or are they analyzed outside a federal constitutional framework, and under a special standard of review, because
individuals under supervision have “compromised” constitutional rights? Does the
analysis change depending on whether the abridged right is fundamental?
There is no consensus over how supervision conditions are analyzed, even
when a statutory framework exists to provide guidance on how conditions should
be structured.52 An in-depth study in 1999 of First Amendment challenges to
“Scarlet-Letter” conditions of probation noted the “chaotic hodgepodge of different standards” for assessing the legality of these conditions.53 Not much has
changed since. Circuits have different standards for assessing conditions that
touch on fundamental rights. Courts review state supervision conditions differently from federal supervised release conditions. Legislatively mandated conditions, codified in state law, might receive a different type of scrutiny from discretionary, judicially imposed conditions.54
The Supreme Court’s decision in Packingham is a helpful lesson in the messiness of First Amendment law and the futility of relying too heavily on neat, doctrinal boxes. The opinion not only passes on the question of which tier of scrutiny
applies to the challenged North Carolina statute, but does not conduct any methodical First Amendment analysis, ultimately likening the case to a prior First Amendment decision involving the overbreadth doctrine, another potential mode of challenging supervision conditions.55 In light of this murky doctrine, rather than
52. See 18 U.S.C. § 3553; 18 U.S.C. § 3583.
53. Phaedra Athena O’Hara Kelly, The Ideology of Shame: An Analysis of First Amendment
and Eighth Amendment Challenges to Scarlet-Letter Probation Conditions, 77 N.C. L. REV. 783,
838 (1999).
54. Only the latter is technically governed by 18 U.S.C. § 3553 and § 3583, the statute setting
the boundaries of federal supervised release conditions.
55. Packingham v. North Carolina, 137 S. Ct. 1730, 1738 (2017) (citing Bd. of Airport
Comm’rs of Los Angeles v. Jews for Jesus, Inc., 482 U.S. 569 (1987)); see also Washington State
Grange v. Washington State Republican Party, 552 U.S. 442, 449 n.6 (2008) (identifying a regulation
as unconstitutionally overbroad when “a substantial number of its applications are unconstitutional,
judged in relation to the statute’s plainly legitimate sweep”) (citation and internal quotation marks
omitted).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

675

identifying a discrete analytical framework for each type of challenge against a
supervision condition, this section distills a handful of principles that would underlie any challenge.
First, the “compromised” nature of supervisees’ constitutional rights does not
give judicial, legislative, or supervisory authorities carte blanche to impose conditions as they see fit. The Supreme Court’s decisions in Morrissey v. Brewer and
Samson v. California have long established that individuals under supervision possess qualified constitutional rights56—just how qualified these rights are, the
Court has never clarified.57 In line with this principle, courts across circuits and
states engage in some form of tailoring for supervision conditions. The question,
therefore, is not whether tailoring occurs, but how much. Some circuits, for example, engage in a higher degree of scrutiny of supervision conditions when the conditions implicate a fundamental right.58 For instance, the Second Circuit has
staked out a more searching form of scrutiny: “If the liberty interest at stake is
fundamental, a deprivation of that liberty is ‘reasonably necessary’ only if the deprivation is narrowly tailored to serve a compelling government interest.”59 This
arguably exceeds the scrutiny required by the federal statute governing supervised
release conditions.60 Other circuits explicitly decline to engage in more searching
scrutiny when parolees and probationers have their fundamental rights burdened.61 Yet underlying these divergent standards is a common requirement that
56. Morrissey v. Brewer, 408 U.S. 471, 482 (1972) (“[T]he liberty of a parolee, although indeterminate, includes many of the core values of unqualified liberty and its termination inflicts a
‘grievous loss’ on the parolee and often on others. . . . By whatever name, the liberty is valuable and
must be seen as within the protection of the Fourteenth Amendment.”); Samson v. California, 547
U.S. 843, 850 n.2 (2006); see also Griffin v. Wisconsin, 483 U.S. 868, 875 (1987) (noting that the
“permissible degree” of state “impingement upon [the] privacy” of individuals under supervision is
“not unlimited”).
57. The Court has acknowledged its own ambiguity around the extent of parolees’ constitutional rights. See Samson, 547 U.S. at 850 n.2 (“In [Morrissey], the Court recognized that restrictions
on a parolee’s liberty are not unqualified. That statement, even if accepted as a truism, sheds no light
on the extent to which a parolee’s constitutional rights are indeed limited—and no one argues that a
parolee’s constitutional rights are not limited.”).
58. See, e.g., United States v. Loy, 237 F.3d 251, 256 (3d Cir. 2001) (“[A] condition that
restricts fundamental rights must be ‘narrowly tailored and . . . directly related to deterring [the defendant] and protecting the public.’”) (citing United States v. Crandon, 173 F.3d 122, 128 (3d Cir.
1999)).
59. United States v. Myers, 426 F.3d 117, 126 (2d Cir. 2005).
60. 18 U.S.C. § 3583(d) (2016) (requiring supervised release conditions to “involve[] no
greater deprivation of liberty than is reasonably necessary” to effectuate the purposes of sentence).
61. See, e.g., United States v. Zinn, 321 F.3d 1084, 1089 (11th Cir. 2003) (“[W]hile the Sentencing Guidelines recognize that a condition of supervised release should not unduly restrict a defendant’s liberty, a condition is not invalid simply because it affects a probationer’s ability to exercise constitutionally protected rights.”); United States v. Schave, 186 F.3d 839, 843 (7th Cir. 1999)
(“[A] court will not strike down conditions of release, even if they implicate fundamental rights, if
such conditions are reasonably related to the ends of rehabilitation and protection of the public from
recidivism.”) (emphasis added).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

676

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

the supervision condition be tailored to the individual, his offense and history, and
sentencing goals.
Second, in cases where the court employs First Amendment analysis, it is
likely that these restrictions will be analyzed as content-neutral rather than content-based. First Amendment analysis may be appropriate in a number of situations. First, when a challenge is directed at a state or local condition, there may be
no statute on point dictating how to assess the challenge. This differs from the
federal context where statutes such as 18 U.S.C. § 3553 and 18 U.S.C. § 3583
guide how to assess the reasonableness of a supervised release condition.62 Second, if a challenge is directed at a mandatory social media ban rather than at a law
granting officers discretion over whether and how to restrict social media access,
the court will likely not invoke a statutory framework, like § 3553 and § 3583,
designed for assessing discretionary conditions.63 If the court analyzes the supervision condition as a constitutional question, the question will be whether it is
content-based, warranting strict scrutiny, or content-neutral, warranting intermediate scrutiny. In the First Amendment universe, a determination of content neutrality is a crucial, threshold win for challenged restrictions facing invalidation—
whereas the application of strict scrutiny almost guarantees invalidation, the application of intermediate scrutiny usually means the opposite.64 This Article is not
the space for determining which tier of scrutiny should apply; it is enough to note
that courts tend to view such restrictions as content-neutral.65
Lastly, both First Amendment intermediate scrutiny and statute-driven scrutiny (via 18 U.S.C. § 3553 and § 3583 or state equivalents) involve narrow tailoring. Intermediate scrutiny requires that a content-neutral regulation be “narrowly
tailored to serve the government’s legitimate, content-neutral interests.”66 And
across circuits, courts discussing the federal statutory guidelines have remarked
that they constitute a “narrow tailoring” requirement.67 Thus, for these general
62. See Mutter v. Ross, 811 S.E.2d 866 (W. Va. 2018) (engaging in First Amendment analysis
of a state supervision condition as a content-neutral regulation of speech). Still, even when courts
are not required to use the federal statutes as guidelines for analyzing supervisions, they may find
them persuasive. See J.I. v. New Jersey State Parole Bd., 155 A.3d 1008, 1022 (N.J. 2017) (explaining that the state court could “gain insight” from federal court cases involving the application of 18
U.S.C. § 3583).
63. Cf. 18 U.S.C. §§ 3553, 3583 (addressing the appropriateness of individualized sentencing).
64. See Barry P. McDonald, Speech and Distrust: Rethinking the Content Approach to Protecting the Freedom of Expression, 81 NOTRE DAME L. REV. 1347, 1351 (2006) (explaining through
empirical analysis that “it is the initial content characterization of a regulation that does all of the
work in basic free speech cases, and if it is labeled as ‘content-based’ it is categorically invalidated,
and if ‘content-neutral’ it is almost categorically upheld”).
65. The Supreme Court’s hesitance to state the proper tier of scrutiny in Packingham is likely
to encourage this view, see infra note 161.
66. Ward v. Rock Against Racism, 491 U.S. 781, 798 (1989).
67. See, e.g., United States v. Duke, 788 F.3d 392, 398 (5th Cir. 2015) (“[T]he condition must
be narrowly tailored such that it does not involve a ‘greater deprivation of liberty than is reasonably

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

677

purposes, judicial review of supervision condition will involve asking the question: Is this condition narrowly tailored to support the government’s interests? The
hodgepodge of analytical standards in case law all coalesce around this basic question.
And it truly is a hodgepodge of standards. For instance, in Mutter v. Ross, the
Supreme Court of Appeals of West Virginia writes that a supervision condition
must be “narrowly tailored so as not to burden more speech than is necessary to
further the government’s legitimate interests,” but never explicitly states that First
Amendment analysis applies.68 And in J.I v. New Jersey State Parole Board, the
Supreme Court of New Jersey fashions a reasonableness analysis that draws on
Third Circuit case law but never quite states the proper standard for assessing
whether a condition is permissible.69 These cases and others lack doctrinal consistency, but they demonstrate that some form of narrow tailoring to further a legitimate governmental interest is necessary, regardless of whether the court uses
First Amendment or statutory analysis. And, as described below, Internet and social media bans will often fail to meet this moderate standard.
B. Conditions on the Wrong People
The first major constitutional flaw with some Internet and social media bans
is their indiscriminate application. While restrictions on Internet and social media
access may be appropriate for actors who commit the most sophisticated, manipulative offenses, the class of people deemed “sex offenders” is simply too diverse
for one-size-fits-all bans. Consider the following hypothetical scenarios, all of
which involve the commission of a sex offense under state laws that ban “sex
offender supervisees” from social media:
1.
A 17-year-old exposes his genitalia in public on multiple occasions.70
2.
A 19-year old forces his 16-year-old girlfriend to perform oral sex
on him.71 It is his first offense.
3.
A 30-year old man corners his 35-year old ex-girlfriend at a party
and improperly touches her intimate parts without her consent.72 He
necessary’ to fulfill the purposes set forth in § 3553(a).”); United States v. Voelker, 489 F.3d 139,
146 (3d Cir. 2007) (“[A]ny such restriction had to be narrowly tailored and consistent with the sentencing factors set forth in 18 U.S.C. § 3553(a).”); United States v. Holm, 326 F.3d 872, 877 (7th
Cir. 2003) (“[T]o the extent that the condition is intended to be a total ban on Internet use, it sweeps
more broadly and imposes a greater deprivation on Holm’s liberty than is necessary, and thus fails
to satisfy the narrow tailoring requirement of § 3583(d)(2).”).
68. 811 S.E.2d 866, 872 (W. Va. 2018).
69. 155 A.3d 1008, 1021–23 (N.J. 2017).
70. See S.C. CODE ANN. § 23-3-430(C)(14) (2018) (indecent exposure).
71. See TEX. PENAL CODE ANN. § 22.011 (West 2018) (sexual assault).
72. See N.Y. PENAL LAW § 130.52 (McKinney 2017) (forcible touching).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

678

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

has an extensive criminal record for groping and improper touching.
4.
A 50-year old man drugs a 20-year old girl at a bar, takes her home
and sexually assaults her.73 It is his first sex offense, though he has
been charged with violent crimes before.
5.
A 52-year old man engages in a sexual relationship with his consenting 28-year old biological daughter.74 It is his first sex offense,
though he has been convicted of tax fraud before.
6.
A 40-year old man buys several photographs on a child pornography
website.75 It is his second offense of this type.
7.
A 38-year old man, pretending to be 25, “friends” a 15-year old on
Facebook and asks her to meet him at a park, where she agrees to
have sexual intercourse with him.76 It is his first offense.
These scenarios are diverse in almost every way imaginable: some involve
direct contact between actor and victim, some only involve remote contact; some
involve adolescent victims, some involve adult victims; some involve predatory
use of the Internet, some do not (some don’t involve the Internet at all); some
involve a preexisting relationship between actor and victim, some involve two
strangers; some involve an actor with an extensive criminal history of similar
crimes, some involve an actor with no record at all.
The diversity of these scenarios, all involving the commission of a sex offense, suggests that a diversity of methods would be effective in furthering the
rehabilitative, deterrent, incapacitory, and (in some cases) punitive goals of supervision.77 For instance, an individual who drugs and violently assaults another
might be barred from visiting establishments which serve alcohol and from acquiring firearms or other weapons, while an individual who extorts a minor for
sexually suggestive photos might be barred from contacting minors on Internet
websites. The recognition that not all individuals who commit sex offenses are the
same should manifest itself in a judge or supervision officer’s discretion in imposing different conditions on different people.
The research supports this intuition. Academic studies of individuals who
commit sex offenses suggest that the crime is not generic and that a predilection

73. See 720 ILL. COMP. STAT. ANN. 5/11-1.50(a)(2) (2017) (criminal sexual abuse via lack of
consent).
74. See 720 ILL. COMP. STAT. ANN. 5/11-11 (2017) (sexual relations within families).
75. See NEV. REV. STAT. § 200.730 (2017) (possession of visual presentation depicting sexual
conduct of person under 16 years of age).
76. See S.C. CODE ANN. § 16-3-655(C) (2015) (criminal sexual conduct with a minor in the
third degree).
77. For an expansive formulation of diverse scenarios involving crime and the correspondingly
diverse purposes in sentencing, see Michael Tonry, Intermediate Sanctions in Sentencing Guidelines, 23 CRIME & JUST. 199, 247–48 (1998).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

679

to reoffend is not uniformly permanent.78 If it were, there might be a rational basis
for treating similarly the first-time, juvenile offender whose sexual misconduct
involved a same-aged cousin and the serial, adult offender whose crimes are highly
sophisticated, extortive ploys perpetrated on young children. But the research
shows otherwise. The most recent scientific study of re-offense, for example,
found that “there is no evidence that individuals who have committed such offenses inevitably present a lifelong enduring risk of sexual recidivism.”79 This
scholarship suggests that the first-time juvenile offender should not be treated as
though they are just a few years away from serially committing manipulative sex
crimes, particularly considering that criminogenic factors like emotional instability and lack of conscientiousness generally decline with age.80 If anything, what
most “sex offenders” share is a general trend to desist from sexual crimes over
time.81 As Professor Eric Janus has remarked, “[m]ost researchers agree . . . that
sexual offending is complex and heterogeneous and that it has multiple independent causes.”82 There is no known “sex offender genome” common to those who
have committed the most minor and the most serious offenses.
By categorizing all the offenses in the list above as undifferentiated “sex offenses,” some states treat all individuals who commit a sex offense as though they
are highly sophisticated, online extortionists.83 That is, these states ban these individuals wholesale from the use of social media or the Internet, simply because
they have once been convicted of a sex offense and now serve a term of supervision. These conditions create a mismatch between crime and condition. Consensual incest may be viewed as morally reprehensible, but does barring its

78. This has not stopped courts across the country—including the Supreme Court—from regurgitating bunk statistics about sex offender recidivism, as Ira and Tara Ellman illuminate in
“Frightening and High”: The Supreme Court’s Crucial Mistake About Sex Crime Statistics, 30
CONST. COMMENT. 495, 499 (2015).
79. R. Karl Hanson, Andrew J. R. Harris, Elizabeth Letourneau, Leslie Maaike Helmus, &
David Thornton, Reductions in Risk Based on Time Offense-Free in the Community: Once a Sexual
Offender, Not Always a Sexual Offender, 24 PSYCHOL. PUB. POL’Y & L. 48, 59 (2018).
80. See Brent W. Roberts, Kate E. Walton & Wolfgang Viechtbauer, Patterns of Mean-Level
Change in Personality Traits Across the Life Course: A Meta-Analysis of Longitudinal Studies, 132
PSYCHOL. BULL. 1, 14–17 (2006); R. Karl Hanson & Kelly E. Morton-Bourgon, The Accuracy of
Recidivism Risk Assessments for Sexual Offenders: A Meta-Analysis of 118 Prediction Studies, 21
PSYCHOL. ASSESSMENT 1, 1–2 (2009).
81. See Hanson et al, supra note 79 at 57–58.
82. Eric S. Janus, FAILURE TO PROTECT: AMERICA’S SEXUAL PREDATOR LAWS AND THE RISE OF
THE PREVENTIVE STATE 51 (2006). See also Eric J. Chan, Dale E. McNiel, & Renee L. Binder, Sex
Offenders in the Digital Age, 44 J. AM. ACAD. PSYCHIATRY LAW 368, 373 (2016) (discussing data
illuminating “different types of sex offenders”).
83. See David T. Goldberg & Emily R. Zhang, Our Fellow American, the Registered Sex Offender, 2016 CATO SUP. CT. REV. 59, 74–75 (“[H]ow did we get to the point at which it is permissible
to view every person on a registry as if he were like the defendant in the 1997 case of Kansas v.
Hendricks, a member of the truly tiny class of persons whose personality disorder compels them to
commit sexual acts against children?”).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

680

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

perpetrators from using the Internet prevent more incest? Is a 17-year-old who
repeatedly exposes himself in public inherently likely to lure unsuspecting minors
into manipulative sexual activity, such that we should bar him from using
LinkedIn to find his first job? This mismatch—conditions that are simply unrelated to the defendant, his offense and history, and the goals of supervision—represents the first infirmity of supervision conditions restricting Internet and social
media access. In the words of the federal statute governing supervised release,
conditions are improper if they are not “reasonably related” to the underlying offense, the history and characteristics of the defendant, and sentencing-related
goals.84 Such mismatched conditions arise via statutory mandate85 and parole officer discretion alike.86 And these mismatches exist despite legal consensus
against blunt categorizations in supervision sentencing.
“Consensus” is not an overstatement, at least in the federal courts. As a Federal Judicial Center report notes, appeals courts typically “caution sentencing
courts not to apply set packages of special conditions to entire classes or categories
of defendants (e.g., all ‘sex offenders’).”87 U.S. Sentencing Guidelines include
special conditions of probation and supervised release that restrict computer and
Internet access “in cases in which the defendant used such items.”88 And across
circuits, in as-applied challenges to Internet restrictions, courts hold that such conditions are impermissible for individuals whose crimes did not involve the Internet
or children. Thus, when courts refer to a “split” in this jurisprudence, they are not
referring to a debate over whether it is permissible to ban social media for those
whose crimes were “offline,”89 or were committed against adults, or were not
84. See 18 U.S.C. § 3583(d)(1) (2012) (citing §§ 3553(a)(1), (a)(2)(B)–(D)).
85. See, e.g., TEX. GOV’T CODE ANN. § 508.1861 (West 2018) (requiring ban on “commercial
social networking site” access for any parolee convicted of a sex offense and assigned a risk level of
two or three, regardless of crime committed and criminal history).
86. See, e.g., Mutter v. Ross, 811 S.E.2d 866, 868 (W. Va. 2018) (parole officer imposed a
ban on Internet access on defendant, who had sexually assaulted an adult woman, a crime not involving the Internet or children, preventing him from receiving emails from an employer or medical
professional, paying a bill online, checking the weather, or using a smartphone).
87. Vance, supra note 29, at 2; see also 8E GUIDE TO JUDICIARY POLICY § 240(d) (“When
considering special condition recommendations, officers should avoid presumptions or the use of set
packages of conditions for groups of offenders and keep in mind that the purposes vary depending
on the type of supervision. Ask first whether the circumstances in this case require such a deprivation
of liberty or property to accomplish the relevant sentencing purposes at this time.”) (emphasis in
original).
88. U.S.S.G. §§ 5B1.3(d)(7)(B), p.s. & 5D1.3(d)(7)(B), p.s.
89. Some courts have likened Internet restrictions to hypothetical restrictions on other methods
of communication related to the crime. See United States v. Peterson, 248 F.3d 79, 83 (2d Cir. 2001)
(“Although a defendant might use the telephone to commit fraud, this would not justify a condition
of probation that includes an absolute bar on the use of telephones.”); see also Ariana Deskins, Internet Use and Sex-Crimes Convicts: Preserving the First Amendment Rights of Sexual Offenders
Through the Framework of United States v. Albertson, 91 U. DET. MERCY L. REV. 29, 45–46 (2014).
This analogy may stretch too far. The Internet allows stalking, luring, and manipulating in ways
wholly unlike what can be done with a telephone. Still, just because the Internet can facilitate such

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

681

predatory at all. On that point, there is a clear consensus: it is not.90 Rather, this
“split” refers to a narrower disagreement over whether possession (as opposed to
creation or distribution) of child pornography is sufficient to justify a broad restriction on the defendant.91 On the more basic point, courts agree: banning a person from the Internet should, in most cases, require evidence that the person will
likely use the Internet to recidivate.
Yet in the face of this consensus, state restrictions bearing no relation to supervisees’ crimes or histories are routinely—and sometimes automatically—imposed. In New York, New Jersey, South Carolina, Texas, Illinois, and Nevada,
state law requires sentencing courts to prohibit individuals under supervision for
sex offenses from using social media.92 While the laws in these six states vary in
form,93 they all ban social media or Internet access for broad categories of “sex
offenders” without regard for individualized assessments of the defendants. In
these states, many of the hypothetical situations listed at the beginning of this section would result in the offender being banned from the Internet or social media
while under supervision. Beyond these statutorily-imposed, class-wide restrictions, individual discretionary conditions banning social media access are routinely imposed on individuals whose crimes had nothing to do with the Internet.
Without a centralized reporter of discretionary conditions imposed across the
country, it is difficult to determine how often this occurs. Various challenges in
case law, however, confirm that legislatively mandated conditions are not the only
kind to restrict Internet access for those whose crimes were offline. For example,
the defendant in Mutter v. Ross, who had previously sexually assaulted an adult in
predation, this does not mean all Internet use prior or integral to a crime meaningfully relates to the
crime itself: an individual who uses directions on Google Maps to find his way to ex-girlfriend’s
house before sexually assaulting her has hardly exhibited a tendency toward predatory Internet practices.
90. See United States v. Holm, 326 F.3d 872, 878 (7th Cir. 2003) (“[T]he decisions of our
sister circuits . . . have also declined to uphold a total ban on Internet access by defendants convicted
of receiving child pornography without at least some evidence of the defendant’s own outbound use
of the Internet to initiate and facilitate victimization of children.”); United States v. Love, 593 F.3d
1, 12 (D.C. Cir. 2010) (noting that “[c]onsensus is emerging among our sister circuits” that Internet
bans are “unreasonably broad for defendants who possess or distribute child pornography”).
91. See Brant, supra note 13, at 781, 785 (discussing the “circuit split” that has formed between
the Fourth, Fifth, Eighth, Ninth, and Eleventh Circuits supporting complete prohibitions, and the
Second, Third, Seventh, and Tenth Circuits preferring “restricting a sex offender’s access to specific
websites”).
92. See N.Y. EXEC. LAW § 259-c(15) (McKinney 2018) (parole); N.Y. PENAL LAW § 65.10 4a (b) (McKinney 2010) (probation); S.C. CODE ANN. § 23-3-555 (2019) (parole); TEX. GOV’T CODE
ANN. § 508.1861 (West 2018) (parole only); 730 ILL. COMP. STAT. ANN. 5/5-6-3.1 (2019) (“supervision”); NEV. REV. STAT. ANN. § 213.1245(p) (West 2017) (parole); Nev. Rev. Stat. Ann. §
176A.410 (probation); N.J. STAT. ANN. § 2C:43-6.6(a) (“sex offenders,” prior approval exception for
parolees and probationers). For a complete table of jurisdictions with statutes limiting sex offenders’
social networking site use as of 2016, see Chan et al., supra note 81, at 370.
93. See supra notes 24, 25, and 26 (describing the variations in these mandatory restriction
statutes).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

682

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

her apartment and stolen money from her purse, was prohibited by his parole office from “possessing or having contact with any computer, electronic device,
communication device or any device which is enabled with internet access.”94 The
defendant was arrested and returned to prison for living with his girlfriend who
owned a computer with Internet access, despite an absence of evidence that he
ever used her password-protected computer.95
Regulations imposing certain social media restrictions on those who have
been proven to use the Internet to engage in child sex abuse may be appropriate to
further the government’s interest in stopping such abuse and in supporting the individual’s rehabilitation and reintegration in society. A regulation burdening a supervisee whose crime did not involve use of the Internet, however, will rarely be
so justified. The weight of the government’s interest does not justify targeting individuals whose incapacitation does nothing to further this interest; that is, the
degree of First Amendment tailoring required does not hinge on how compelling
the underlying interest is.96 This is a basic precept of constitutional law. Narrow
tailoring, whether toward a significant or a compelling governmental interest, requires a closer fit between crime and condition.
C. Conditions on the Wrong Speech
Suppose, however, that the fit between crime and condition is not too attenuated. For instance, a restriction on Internet access for an individual whose crime
involved the Internet, or who displayed a proclivity for soliciting vulnerable minors through some technological means. This is the context where the commitment
to narrow tailoring matters most, where the depravity of a crime cannot trump
neutral legal principles.
Assuming that a prohibition on Internet or social media amounts to prohibiting speech,97 such a prohibition must be narrowly tailored to further a governmental interest. In the federal supervised release context, courts conduct this narrow
tailoring under the umbrella of 18 U.S.C. § 3583(d)(2), which prohibits a “greater
deprivation of liberty than is reasonably necessary” to effectuate the purposes of
sentencing. As noted in Part II.A, whether a court uses intermediate scrutiny, 18
U.S.C. § 3583, or another analytical framework to assess a supervision condition,
the question will be whether the speech abridgement is narrowly tailored.
Such prohibitions can be overbroad, or fail narrow tailoring, for three reasons.
First, a substantial amount of protected speech is silenced by these restrictions.

94. Mutter v. Ross, 811 S.E.2d 866, 868 (W. Va. 2018) (alterations omitted, emphasis in original).
95. Id. at 869.
96. See Goldberg & Zhang, supra note 83, at 59, 66 (“It might understandably be interjected
that these rules can’t really apply when the harm targeted is serious—preventing litter is one thing,
but sexual abuse of a minor is another. The Court’s precedents have this answer: ‘No.’”).
97. After Packingham, this is no longer just an assumption. See infra Part III.C.1.

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

683

This includes, for example, political speech, speech related to economic activity,
and speech related to interpersonal relationships. Second, there are not sufficient
offline substitutes for what happens online. Third, there are many less restrictive
means available to the government besides intrusive bans.
1.

The Range of Speech Burdened by Internet and Social Media
Restrictions

The burdens on protected speech are not incidental to Internet and social media bans—they are the crux of these bans. This section highlights speech activity98
burdened by Internet and social media restrictions and its relevance to the supervision context for those convicted of sex offenses.
Political speech. Restrictions on Internet and social media access necessarily
inhibit “speaking and listening in the modern public square,” as the Supreme Court
held in Packingham.99 Indeed, at least two types of political speech are hampered
by such restrictions. First, political speech “posted” on a social media platform is
shut down. This is speech directed at contemporaries, or at no one in particular,
for the purpose of discussion on matters of public concern. Second, speech petitioning political representatives to take action on a certain issue is silenced. As
representatives increasingly engage with their constituents via social media, such
a bar removes a primary mode of political speech.100 Furthermore, beyond these
two restrictions on speaking, the ability to receive information about how to participate in civic life is severely constrained without Internet or social media access.
Board of elections websites, advertisements from political parties, primers from
interest groups on political candidates—all of these sources of information are
unavailable to the parolee or probationer blocked from the Internet and social media. These restrictions on political speech are not only unnecessary for the purposes of incapacitation and deterrence, but they doom efforts to reintegrate supervisees as active participants in civil society.
Speech involving economic activity. A supervision condition banning social
media twenty years ago would not have posed too difficult a barrier to employment. Even ten years ago, the supervisee’s most fruitful avenue of finding employment was “the purchase of a newspaper or an employment guide.”101 But this
98. As a reminder, I use the term “speech” to include both speaking and receiving information.
See supra note 18.
99. Packingham v. North Carolina, 137 S. Ct. 1730, 1737 (2017).
100. See Knight First Amendment Inst. v. Trump, 302 F. Supp. 3d 541, 549 (S.D.N.Y. 2018)
(“This case requires us to consider whether a public official may, consistent with the First Amendment, ‘block’ a person from his Twitter account in response to the political views that person has
expressed, and whether the analysis differs because that public official is the President of the United
States. The answer to both questions is no.”).
101. Wendy Heller, Poverty: The Most Challenging Condition of Prisoner Release, 13 GEO.
J. ON POVERTY L. & POL’Y 219, 233 (2006).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

684

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

reality has changed drastically, with the Internet and social media now the primary
job-seeking tool for those in need of employment.102 Overbroad restrictions on
the use of social media also create basic burdens on consumer activity. For instance, in the realm of e-commerce, Amazon and other consumer websites may be
classified as social media, given users’ ability to create personal profiles, interact
with other users, leave comments, and respond to comments.103 A supervisee
whose access restrictions are defined to include such websites is thus blocked from
online shopping. Restrictions on social media also prevent covered individuals
from effectively promoting their own businesses. This is particularly burdensome
for individuals who are already roadblocked from entering traditional job markets.104 Beyond the usual burdens facing individuals convicted of felonies, those
convicted of sex offenses are further prohibited by law from working in certain
geographic areas105 and with certain clientele,106 never mind the “significant evidence of onerous practical effects of being listed on a sex offender registry,” as
Justice Souter put it.107 Entrepreneurship is one way to create business opportunities outside of these restrictions, and social media is a vital tool—perhaps the most
vital tool—for becoming a successful entrepreneur.108 It is difficult to imagine a
102. See United States v. Eaglin, 913 F.3d 88, 96 (2d Cir. 2019) (“[O]ne of the conditions of
supervised release is that [the defendant] remain employed: to search for a job in 2019, the Internet
is nearly essential, as the Court in Packingham recognized.”); Amicus Curiae Brief of Electronic
Frontier Foundation et al. in Support of Petitioner at 21–23, Packingham v. North Carolina, 137 S.
Ct. 1730 (2017) (No. 15-1194) (canvassing the various ways working Americans rely on social media for employment, networking, and finding resources necessary for job performance).
103. Packingham, 137 S. Ct. at 1736 (“[G]iven the broad wording of the North Carolina statute
at issue, it might well bar access not only to commonplace social media websites but also to websites
as varied as Amazon.com, Washingtonpost.com, and Webmd.com.”); id. at 1741 (Alito, J., concurring).
104. See Elena Saxonhouse, Unequal Protection: Comparing Former Felons’ Challenges to
Disenfranchisement and Employment Discrimination, 56 STAN. L. REV. 1597, 1610–14 (2004)
(highlighting social and legal consequences of conviction on employment); see generally Stacy A.
Hickox, A Call to Reform State Restrictions on Hiring of Ex-Offenders, 12 STAN. J. CIV. RTS. & CIV.
LIBERTIES 121 (2016).
105. Lester Packingham himself had been forced to quit a job in a shopping mall kiosk because
there was a daycare facility on the premises. Goldberg & Zhang, supra note 83, at 74.
106. The Justice Center of the Council of State Governments compiles the National Inventory
of the Collateral Consequences of Conviction. NATIONAL INVENTORY OF THE COLLATERAL
CONSEQUENCES OF CONVICTION, https://niccc.csgjusticecenter.org [https://perma.cc/5CS2-SWYS]
(last visited Jan. 28, 2019). The inventory specifies the various employment restrictions placed on
individuals convicted of sex offenses, ranging from the expected (elementary school teachers) to the
“really?” (ice cream truck drivers in Massachusetts). Id.
107. Smith v. Doe, 538 U.S. 84, 109 (2003) (Souter, J., concurring).
108. This was the case for Sherman Manning who turned to Baptist preaching as a vocation
after his release from prison. As noted above, Mr. Manning was barred by parole from engaging in
the primary form of promoting his professional services: posting YouTube videos of his sermons.
Manning v. Powers, 281 F. Supp. 3d 953, 957 (C.D. Cal. 2017). The ban on social media effectively
foreclosed Mr. Manning’s ability to advertise his craft, which a federal district court found likely to
violate the First Amendment. Id. at 966.

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

685

new business successfully promoting itself only with newspaper advertisements
and flyers around the neighborhood. Yet again, such supervision conditions impede the very reintegration they are supposed to further.
Speech involving interpersonal relationships. A ban on Internet and social
media shuts down an integral tool of communication between covered individuals
and those dear to them. Of primary significance to many individuals subject to
broad restrictions is their inability to contact family members and friends on social
media. As the Ninth Circuit stated in a pre-Packingham supervision condition
case, “[u]se of the Internet is vital for a wide range of routine activities in today’s
world . . . [such as] communicating with friends and family, . . . [c]utting off all
access to the Internet constrains a defendant’s freedom in ways that make it difficult to participate fully in society. . . .”109 Supervisees who seek progress, closure,
forgiveness, love, and other things from those close to them need the most basic
technological tools—a Facebook chat, an email, a direct message on Instagram—
to do so. There are also new relationships, including romantic ones, to be formed.
Yet some restrictions exclude supervisees from mobile dating applications, used
by a rapidly increasing percentage of Americans, even if the mobile application
itself forbids minors from using it and does not exclude those required to register
as “sex offenders” from becoming members. Hinge and Grindr, for example, do
not prohibit individuals convicted of sex offenses from creating profiles on their
platforms, but may be off limits for a supervisee barred from the use of social
media.110
The breadth of silenced speech in each of these categories is all the more
striking when one recalls that preventing more crime is not the only object of supervision, nor is supervision driven primarily—or at all, in the case of federal supervised release—by a punitive rationale. Rather, supervision conditions are supposed to burden no more liberty than reasonably necessary to further the interests
of rehabilitation and reintegration. To be sure, a court is not obligated to refer to
each of the various purposes of supervision every time it imposes a condition. As
congressional reports from the enactment of sentencing legislation explain, some
cases will simply not involve one or more purposes of sentencing: “In setting out
the four purposes of sentencing, the Committee has deliberately not shown a preference for one purpose of sentencing over another in the belief that different purposes may play greater or lesser roles in sentencing for different types of offenses
109. United States v. LaCoste, 821 F.3d 1187, 1191 (9th Cir. 2016).
110. See HINGE, https://hinge.co/terms/ [https://perma.cc/H7DJ-86T8] (last updated August 2,
2018) (no prohibition of individuals convicted of sex offenses from joining dating application);
Grindr Terms and Conditions of Service, GRINDR, https://www.grindr.com/terms-of-service/
[https://perma.cc/9RQA-AMSN] (last updated July 1, 2018 ) (same as Hinge). But see Terms of Use,
TINDER, https://www.gotinder.com/terms [https://perma.cc/W9EZ-8UVG] (last updated May 9,
2018) (excluding from eligibility to create an account anyone who is “required to register as a sex
offender”);
Terms
&
Conditions,
OKCUPID,
https://www.okcupid.com/legal/terms
[https://perma.cc/XGM2-W3FJ] (last updated Feb. 12, 2019) (same as Tinder).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

686

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

committed by different types of defendants. The Committee recognizes that a particular purpose of sentencing may play no role in a particular case.”111 But as a
general matter, the non-punitive goals driving supervision make Internet bans not
simply an overzealous means of serving an end, but a wrench in the purposes of
supervision itself.
2. The Lack of Internet Alternatives
A much-debated component of intermediate scrutiny in First Amendment
analysis is the requirement that a content-neutral regulation “leave open ample
alternative channels for communication.”112 The statutory analysis for supervised
release conditions in 18 U.S.C. § 3583 similarly implies that some channels of
communication should be left open to a supervisee, as conditions of release may
involve “no greater deprivation liberty than is reasonably necessary,” as discussed
above. These standards draw on the Supreme Court’s suggestion in City of Ladue
v. Gilleo that a regulation foreclosing one avenue of speech must leave open other,
adequate ways for the speaker to make their point.113 Ladue involved a municipal
ban on lawn signs, which the Court struck down on the grounds that no adequate
alternative existed for this “unusually cheap and convenient” means of expression.114 Here, it is difficult to imagine offline speech serving as an adequate substitute for online speech.
But there are closer substitutes than purely offline speech for social media. In
cautioning against too broad a reading of Ladue, Professor Noah Feldman has
opined, “the real problem [in Ladue] was that there’s a specific social meaning
attached to putting up a political sign in your front yard: It’s speech that uniquely
is associated with you. But that’s not true of your Facebook page, because in the
absence of a Facebook account you could create your own website with identical
content.”115 This argument could carry some weight if particular social media
sites proscribed had adequate alternatives that didn’t carry the risk of unlawful
activity. For instance, the government could restrict a supervisee’s access to online
teen chatrooms (where the official recommended ages are 13–18), under the theory that any protected speech expounded to willing listeners on TeenChat.com
could just as easily be uttered on Facebook, Twitter, or similar sites. Banning a
supervisee’s access to online teen chatrooms would still allow the supervisee ample online space to engage in protected speech. Feldman’s proposal, however,
takes this idea a step further, suggesting categorically that social media sites do
111. See S. REP. NO. 98-225, at 77 (1983), reprinted in 1984 U.S.C.C.A.N. 3183, 3260.
112. Ward v. Rock Against Racism, 491 U.S. 781, 791 (1989).
113. 512 U.S. 43, 54 (1994).
114. Id. at 57.
115. Noah Feldman, Sex Offenders Don’t Have a Right to Facebook, BLOOMBERG (May 12,
2016, 2:06 PM), https://www.bloomberg.com/view/articles/2016-05-12/sex-offenders-don-t-havea-right-to-facebook [https://perma.cc/R46Q-9U96].

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

687

not have specific social meaning, and that speech on personal websites represents
an adequate alternative to posting on Facebook or tweeting on Twitter.
Anyone who has burrowed down the rabbit holes of Facebook comment wars
or Twitter threads knows this to be incorrect—and the courts are catching on. The
ability to instantly share and respond to content—or to have one’s content instantly
shared and responded to—is the interactivity feature of social media websites
which a personally-owned website cannot replicate.116 This interactivity feature
has gained currency in the ongoing litigation over President Donald Trump’s ability to block Twitter users who have criticized him. In her partial ruling against
President Trump and White House Social Media Director, Dan Scavino, U.S. District Judge Naomi Reice Buchwald of the Southern District of New York highlighted the “‘interactive space’ associated with each of the President’s tweets” as
the forum to which the users sought access.117 Unlike a Twitter timeline or the
mere content of the tweets, which by themselves are mere aggregates of government speech and function identically on www.donaldtrump.com, “the essential
function of a given tweet’s interactive space is to allow private speakers to engage
with the content of the tweet.”118 Although the district court’s discussion is situated in forum analysis, its recognition of the singularity of social media for First
Amendment expression is applicable here. Even if other “online” avenues exist
for protected speech when one’s social media access is shut down, these avenues
are inadequate substitutes for the expressive interactivity which one finds on Facebook, Twitter, and the like, but not on do-it-yourself personal websites.
3. Other Means of Furthering the Governmental Interest
Assuming content-neutral analysis, the federal sentencing guidelines, or some
other less-than-strict-scrutiny framework applies to a supervision condition challenge, the government will not be required to demonstrate that a supervision condition is the least restrictive means of furthering its interest.119 But alternative
means of furthering the government’s interest are still relevant in the context of
116. For another response to Professor Feldman’s theory, see David Post, On ‘Ample Alternative Channels of Communication,’ the First Amendment, and Social Networking, WASH. POST (May
16, 2016), https://www.washingtonpost.com/news/volokh-conspiracy/wp/2016/05/16/on-ample-alternative-channels-of-communication-the-first-amendment-and-social-networking/
[https://perma.cc/GUU8-ERVF].
117. Knight First Amendment Inst. v. Trump, 302 F. Supp. 3d 541, 573 (S.D.N.Y. 2018).
118. Id.; see also id. at 575 (“The interactivity of Twitter is one of its defining characteristics,
and indeed, the interactive space of the President’s tweets accommodates a substantial body of expressive activity.”).
119. Ward v. Rock Against Racism, 491 U.S. 781, 798 (1989) (clarifying that a narrowly
tailored, content-neutral regulation “need not be the least restrictive or least intrusive means” of
serving the government’s interest). But see Vance, supra note 29, at 3 (“Under Judicial Conference
policy, the specific blend of supervision interventions selected by federal probation officers should
be the least restrictive necessary to meet the objectives of supervision in the individual case.”).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

688

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

intermediate scrutiny. As the Supreme Court noted in assessing a content-neutral
statute in McCullen v. Coakley, “[t]o meet the requirement of narrow tailoring, the
government must demonstrate that alternative measures that burden substantially
less speech would fail to achieve the government’s interests, not simply that the
chosen route is easier.”120 But as courts invalidating absolute bans on Internet
access commonly note, the presence of other available means of furthering the
government’s interest is a good sign that the regulation is not narrowly tailored.121
Based on this premise, the compromised nature of supervisees’ Fourth
Amendment rights currently enables the government to monitor their activity in
ways less restrictive of their First Amendment rights. Put simply, court-ordered
surveillance of supervisees limits the need for blunt proscription of all online
speech. This is not to say that supervisees’ watered-down privacy protections are
as robust as they should be, nor that even the limited protections they possess are
respected by law enforcement officers.122 It just means that weak Fourth Amendment protections for this class of people diminishes the need to limit their First
Amendment protections.
This is not the usual story of increased technological precision intersecting
with civil liberties. Civil libertarians often perceive advances in technology as a
threat to privacy in particular. In the context of policing and the Fourth Amendment, this manifests itself in concerns over searches of cell-site location information,123 GPS tracking on cars,124 device searches of cell phones at the United
States border,125 and more. Concern over increased technological precision also

120. McCullen v. Coakley, 134 S. Ct. 2518, 2540 (2014).
121. In Packingham v. North Carolina, after noting that the opinion “should not be interpreted
as barring a State from enacting more specific laws than the one at issue[,]” Justice Kennedy remarks: “Though the issue is not before the Court, it can be assumed that the First Amendment permits a State to enact specific, narrowly tailored laws that prohibit a sex offender from engaging in
conduct that often presages a sexual crime, like contacting a minor or using a website to gather
information about a minor.” 137 S. Ct. 1730, 1737 (2017). Such laws “must be the State’s first resort
to ward off the serious harms that sexual crimes inflict.” Id. This is far from a least restrictive means
requirement, but it clearly demands that the State hew regulations to a significant interest.
122. For critiques of the common requirement that supervisees’ sign away privacy rights, see
William R. Rapson, Extending Search-and-Seizure Protection to Parolees in California, 22 STAN.
L. REV. 129, 140 (1969) (“Requiring normal probable cause for general parolee searches and seizures
and variable probable cause in exceptional cases offers substantial security against criminal activity
without the debilitating effects on rehabilitation accompanying the Hernandez rule. The phrase ‘parole officer’ should not operate as a talisman to legitimate a search or seizure.”); Taylor S. Rothman,
Fourth Amendment Rights of Probationers: The Lack of Explicit Probation Conditions and Warrantless Searches, 2016 U. CHI. LEGAL F. 839, 867 (2016) (“[T]he idea that rehabilitative goals of
probationers are impeded by a warrant requirement is erroneous. In fact, indiscriminate searches
could undermine the rehabilitative process.”).
123. See Carpenter v. United States, 138 S. Ct. 2206, 2210–2211 (2018).
124. See United States v. Jones, 565 U.S. 400, 430 (2012).
125. See United States v. Vergara, 884 F.3d 1309, 1312 (11th Cir. 2018).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

689

extends to other civil liberty realms, such as the field of reproductive rights.126
In this context, by contrast, as law enforcement technology becomes more
precise, the need for imprecise tools—like wholesale Internet bans—starts to
sound fishy. Without an effective, narrowly tailored means of monitoring and restricting the online activities of those deemed likely to commit sex crimes using
the Internet, social media bans may appear harsh to reviewing courts, yet nevertheless be viewed as the only feasible way of protecting the public. But as technology develops, a ban may be unnecessary given the variety of other supervision
options. Monitoring technology is the main example.127 Rather than forbidding
social media usage in part or in its entirety, why not inform the supervisee that
supervision officers will be able to identify any improper online conduct through
remote monitoring? Some states have enacted legislation providing for this very
option of more precise supervision.128 Other less restrictive means include forbidding certain supervisees from visiting specific websites or barring the use of anonymous web browsing.129 Though some scholars have advocated more robust
Fourth Amendment protections for supervisees, which could render these surveillance conditions invalid, there does not appear to be significant momentum in this
direction.130 Any of these means of maintaining public safety are likely to be upheld, and all of them make wholesale bans on the Internet and social media for
supervisees seem even less necessary.
There is one other widely adopted, yet constitutionally insufficient, way of
meeting the demand for less restrictive means: prior-approval provisions in
126. The law surrounding reproductive rights provides one useful example of the intersection
between technological advancements and legal standards. Given that the Supreme Court has reaffirmed viability as the defining line for when an abortion is constitutionally permissible, technological advancements bringing viability earlier could necessarily allow heightened restrictions on a
women’s right to an abortion. See Planned Parenthood of Southeastern Pennsylvania v. Casey, 505
U.S. 833, 846 (1992) (“Before viability, the State’s interests are not strong enough to support a
prohibition of abortion or the imposition of a substantial obstacle to the woman’s effective right to
elect the procedure.”); I. Glenn Cohen, Artificial Wombs and Abortion Rights, 47 THE HASTINGS
CENTER REPORT (July 27, 2017) (considering the effect of artificial womb development on abortion
rights jurisprudence, or “if fetal transfer becomes an option”).
127. Vance, supra note 29, at 1 (noting federal judges’ ability to authorize “the use of hardware
or software to filter, monitor, or record computer and Internet data”).
128. See, e.g., LA. STAT. ANN. § 15:561.5(16) (conditioning release on a supervisee submitting
his Internet-related activities “to continued supervision, either in person or through remote monitoring”).
129. But see McIntyre v. Ohio Elections Comm’n, 514 U.S. 334, 342 (1995) (“[A]n author’s
decision to remain anonymous, like other decisions concerning omissions or additions to the content
of a publication, is an aspect of the freedom of speech protected by the First Amendment.”).
130. However questionable the premise of stripping supervisees of their Fourth Amendment
rights, courts do not appear ready to undo this doctrine. In a case challenging a Wisconsin law that
required persons released from civil commitment to wear a GPS ankle monitor 24 hours a day for
the rest of their lives, Judge Richard Posner opined in response to the contention that such monitoring
of a person’s movements required a search warrant: “That’s absurd.” Belleau v. Wall, 811 F.3d 929,
936 (7th Cir. 2016).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

690

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

supervision conditions. For example, rather than outright banning the use of social
media, a supervision agreement may include the following provision: “Condition
#[X] requires that [the defendant] seek and obtain approval from [his] probation
officer before using any particular computer or computer-related device, internetservice provider, or computer or internet account—such as a screen user name or
email account.”131 A survey of existing case law on federal supervised release
conditions reveals that prior-approval provisions are widespread.132 But prior-approval clauses, which place a burden on the supervisee to address each potential
violation, do not solve the underlying problem of broadly restricting a primary
medium of speech.133 There are many reasons why the average supervisee would
not seek prior approval for a post, even if the terms of his release permitted this:
supervisees may want to maintain good standing with a supervision officer, and
raising minor issues about social media access might rock the boat. Other supervisees may prioritize raising other grievances over this one. As a result, they would
sacrifice requesting Facebook usage over challenging a residency restriction, for
example—and supervisees may not even realize that they have the ability to challenge a condition. Though there is no legal consensus on the constitutionality of
prior-approval clauses, several courts have refused to say they cure Internet or
social media restrictions of their overbreadth.134
Each of these three components of narrow tailoring—the scope of the speech
burdened, the lack of an alternative to social media, and the availability of means
less restrictive than banning speech—reveal startling constitutional infirmities
with these Internet and social media bans. Given the rehabilitative aim of supervised release from prison, it is difficult to understand how a flat ban on Internet
131. E.g., United States v. Sales, 476 F.3d 732, 736 (9th Cir. 2007).
132. See, e.g., United States v. Dallman, 886 F.3d 1277, 1280 (8th Cir. 2018) (“The defendant
shall not possess or use any computer or electronic device with access to any ‘on-line computer
service’ without the prior approval of the Probation Office.”); United States v. Love, 593 F.3d 1, 11
(D.C. Cir. 2010) (“The defendant shall not possess or use a computer that has access to any ‘on-line
computer service’ at any location, including his place of employment, without the prior written approval of the Probation Office.”); United States v. Ramos, 763 F.3d 45, 51 (1st Cir. 2014) (“Included
in the special conditions of supervision were requirements that Ramos ‘shall not possess or use a
computer that contains an internal, external or wireless modem without the prior approval of the
Court,’ and that he ‘shall not possess or use a computer, cellular telephone, or any other device with
internet accessing capability at any time and/or place without prior approval from the probation officer.’”); United States v. Miller, 594 F.3d 172, 177 (3d Cir. 2010) (“The defendant shall not use a
computer with access to any ‘on-line computer service’ without the prior written approval of the
probation officer.”) (alterations omitted).
133. See Wiest, supra note 14, at 862 (arguing that prior-approval restrictions are punitive,
and thus contrary to the Federal Sentencing Guidelines).
134. See United States v. LaCoste, 821 F.3d 1187, 1192 (9th Cir. 2016) (“The government
seeks to defend the condition as drafted by arguing that it is not really a total ban, since it allows
LaCoste to use the Internet so long as he first obtains his probation officer’s approval. That proviso
does not save what is otherwise a plainly overbroad restriction on LaCoste’s liberty.”); United States
v. Scott, 316 F.3d 733, 734, 737 (7th Cir. 2003) (vacating the “unusual term” in a supervised release
condition that required the defendant to seek prior approval from his probation officer before accessing the Internet).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

691

access helps further this goal, let alone in a targeted fashion.
III.
PACKINGHAM AND ITS BENEFICIARIES
As the discussion above shows, there are grave theoretical and constitutional
issues with Internet and social media bans. And as a practical, strategic matter, the
Supreme Court’s recent decision in Packingham v. North Carolina could—and
should—open the door to parolee and probationer challenges of these bans on a
nearly identical issue.
This Part provides background on the Packingham decision, a defense of its
application in the supervision context, and an explanation of how the decision
strengthens challenges to unconstitutional supervision conditions.
A. The Packingham Decision
In 2008, the North Carolina legislature passed a law making it a crime for any
registered sex offender—whether on supervision or not—to access “a commercial
social networking Web site where the sex offender knows that the site permits
minor children to become members or to create or maintain personal Web
pages[...].”135 When Lester Packingham, who had been a registrant for eight years
after a conviction for taking indecent liberties with a minor when he was in college, posted “Thank you, Jesus!” on Facebook to celebrate getting a traffic ticket
dismissed, he was arrested for violating the law.136
The law was an easy target: It made no distinction between individuals whose
crimes did and did not involve the Internet or who were “high risk” and “low risk”;
it imposed a new criminal penalty on sex offender registrants; and it did not distinguish between individuals still under supervision for their offenses and individuals who had fully completed their sentences. As Mr. Packingham’s counsel wrote
after the case, this was “the kind of law the justices are comfortable striking
down.”137
And comfortably strike it down they did. In an 8-0 opinion authored by Justice
Kennedy, the Court, without deciding whether the North Carolina law was content-based or content-neutral for First Amendment purposes, held that the law
would not even withstand intermediate scrutiny if it were deemed content-neutral.138 The opinion has three short sections, which boil down to the following
propositions: 1) the Internet is an important—even “the most important”—means
of exercising First Amendment rights;139 2) though the North Carolina law serves
135.
136.
137.
138.
139.

N.C. GEN STAT. § 14-202.5 (2009).
Packingham v. North Carolina, 137 S. Ct. 1730, 1734 (2017).
Goldberg & Zhang, supra note 83, at 59, 68.
Packingham, 137 S. Ct. at 1736.
Id. at 1735.

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

692

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

a valid governmental interest, it sweeps far more broadly than necessary, particularly when narrower means are available to serve the interest;140 and 3) no Supreme Court case has ever upheld such a broad restriction of First Amendment
rights.141 Justice Alito, joined by Chief Justice Roberts and Justice Thomas, wrote
separately to state their concerns with “sex offenders” reentering society, and more
substantively, in Part II, to criticize Justice Kennedy for his “loose rhetoric” on
the quintessential public forum status of the Internet.142
B. Who Does Packingham Benefit?
One could read Packingham’s concern over categorical social media restrictions on “sex offenders” and assume that this concern applies broadly to restrictions on sex offender parolees and probationers, who possess clear though
compromised First Amendment rights. Indeed, since the decision, many courts
have either not addressed possible differences between the criminal statute in
North Carolina and supervision conditions they are reviewing143 or have affirmatively stated that these differences don’t matter.144 Others have acknowledged the
differences between these contexts, but have found Packingham relevant to supervision condition challenges.145 Yet several courts addressing challenges to supervision conditions have distinguished Packingham as wholly inapplicable to these

140. Id. at 1736.
141. Id. at 1737.
142. Id. at 1743 (Alito, J., concurring).
143. United States v. Morgan, 696 F. App’x. 309 (9th Cir. 2017) (special condition of supervised release vacated and remanded to the district court in light of Packingham); United States v.
Ely, 705 F. App’x 779, 781 (11th Cir. 2017); United States v. Maxson, 281 F. Supp. 3d 594, 599 (D.
Md. 2017) (noting that Packingham “may have somewhat strengthened Defendant’s position” that
a supervised release condition banning Internet access was overbroad); Manning v. Powers, 281 F.
Supp. 3d 953, 960 (C.D. Cal. 2017); United States v. Avila, No. 17-10065, 719 F. App’x 591, 594
(9th Cir. Dec. 21, 2017) (citing Packingham for the view that a supervised release condition blocking
Internet access “indisputably implicates a significant liberty interest”) (citation and internal quotation marks omitted); State v. Ranstead, No. S-16365, 421 P.3d 15, 20 (Alaska 2018) (citing Packingham for the proposition that “a condition restricting internet access must be narrowly tailored”).
144. See Millard v. Rankin, 265 F. Supp. 3d 1211, 1228 (D. Colo. 2017) (discussing the relative sweep of SORA’s registration requirement and the statute in Packingham); Mutter v. Ross, 811
S.E.2d 866, 872 (W. Va. 2018) (“Packingham made no exception for parolees. Thus, we decline to
accept the State’s argument that Mr. Ross’s status as a parolee, by itself, renders his special condition
of parole constitutional.”); In re Cruz R., No. F073755, 2017 WL 5714088, at *2 (Cal. Ct. App. Nov.
28, 2017) (“Packingham addressed a criminal statute rather than a condition of probation. Nonetheless, as we explain below, we find the Supreme Court’s decision necessarily compels striking the
probation condition here as unconstitutionally overbroad.”).
145. See United States v. Eaglin, 913 F.3d 88, 96 (2d Cir. 2019) (“Certain severe restrictions
may be unconstitutional when cast as a broadly-applicable criminal prohibition, but permissible
when imposed on an individual as a condition of supervised release. In our view, Packingham nevertheless establishes that, in modern society, citizens have a First Amendment right to access the
Internet.”) (citation omitted).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

693

challenges,146 often without much explanation.147 As the Fifth Circuit recently
held, “[b]ecause supervised release is part of [the defendant]’s sentence (rather
than a post-sentence penalty), see 18 U.S.C. § 3583(a), . . . we find that Packingham does not—certainly not ‘plainly’—apply to the supervised-release context.”148 Before delving into the impact of Packingham on supervision conditions,
therefore, it must be resolved whether the opinion is even relevant in this context.149
This question initially arose in back-and-forth party briefings, where the petitioners had a choice to make: distinguish Lester Packingham and other sex offender registrants who were finished with their criminal sentences from those still
serving out their sentences as sex offender supervisees, or rely upon the two
groups’ common protection from burdensome requirements that are not narrowly
tailored. Initially, in their petition for certiorari, the registrant petitioners took the
‘We’ve got it better than they do’ route, distinguishing themselves from sex offender supervisees who resided on a lower constitutional tier.150 In its opposition
146. See United States v. Browder, 866 F.3d 504, 511 n.26 (2d Cir. 2017) (distinguishing
Packingham on the grounds that the North Carolina law at issue “extended beyond the completion
of a sentence”); Weida v. State, 83 N.E.3d 704, 716 n.10 (Ind. App. 2017) (noting that Packingham
was not applicable to a sex offender subject to supervised release); United States v. Pedelahore, No.
1:15cr24-LG-RHW, 2017 WL 4707458, at *2 (S.D. Miss. Oct. 19, 2017) (“The Packingham decision is inapplicable to Pedelahore’s circumstances. Even while on supervised release, Pedelahore is
serving his criminal sentence, and the Court has broad discretion in establishing the conditions under
which Pedelahore will serve the supervised release portion of his sentence.”); United States v. Rock,
863 F.3d 827, 831 (D.C. Cir. 2017) (noting that the challenged condition was “part of [the defendant’s] supervised-release sentence, and is not a post-custodial restriction of the sort imposed on
Packingham”); United States v. Farrell, No. 4:06-CR-103, 2018 WL 1035856, at *2 (E.D. Tex. Feb.
23, 2018).
147. See, e.g., Richardson v. Becerra No. 2:17-cv-01838, 2018 WL 1173820, at *8 (E.D. Cal.
Mar. 6, 2018) (“Packingham . . . is about the ‘relationship between the First Amendment and the
modern Internet’ and not about the validity of laws requiring sex offender registration or publication
of sex offender registries.”).
148. United States v. Halverson, 897 F.3d 645, 658 (5th Cir. 2018).
149. The other main argument used by defenders of state laws and conditions in distinguishing
Packingham is that the North Carolina law at issue targeted “access,” while some challenged laws
and conditions narrowly target “use.” For example, the defenders of Kentucky’s social media use
ban (KRS § 17.546(2)) argued in court that its proscription of “use” as opposed to “access” meant
that a sex offender registrant could still “log onto a website such as Facebook or Twitter and read
the speech of others, thereby obtaining political news, learning of community events, or viewing job
postings.” See Doe v. Kentucky ex rel. Tilley, 283 F. Supp. 3d. 608, 611–613 (E.D. Ky. 2017). The
court rejected this argument, emphasizing that Packingham made explicit reference to speech, rather
than listening, as the cornerstone of First Amendment protections. Id. (“KRS § 17.546 as it currently
stands may allow Mr. Doe to ‘listen’ to the speech of others on social media sites, but it surely does
not allow him to ‘speak.’”).
150. Petition for a Writ of Certiorari at 10, 27, Packingham v. North Carolina, 137 S. Ct. 1730
(2017) (No. 15-1194) (“[P]ersons who are no longer under criminal justice supervision are entitled
to full, not watered-down, First Amendment protections.”) (“[O]ther courts have overturned
measures imposing far less onerous and sweeping burdens upon internet activities of individuals
entitled to less robust constitutional protections (e.g., those still subject to supervised release).”).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

694

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

to certiorari, North Carolina emphasized the petitioners’ concession that those under supervision are “entitled to less robust constitutional protection.”151 Once certiorari had been granted, the petitioners’ merits brief again alluded to different
protections for supervisees,152 but this time affirmed that “even persons under active criminal justice supervision retain First Amendment rights.”153 In all of this
back and forth, apart from the primary question in the case—whether the Court
would strike down the North Carolina law as violative of the First Amendment—
a secondary question lurked: would the Court issue a decision narrowly based on
the law’s application to sex offender registrants who were finished with their criminal sentences?
It did not. The opinion begins by reciting the “fundamental principle of the
First Amendment [] that all persons have access to places where they can speak
and listen,” with no carve-out for those covered by the North Carolina law who
were still under criminal supervision.154 More significantly, when Justice Kennedy pauses to specify who can benefit from access to social media, he writes:
“Even convicted criminals—and in some instances especially convicted criminals—might receive legitimate benefits from these means for access to the world
of ideas, in particular if they seek to reform and to pursue lawful and rewarding
lives.”155 “Convicted criminals” include individuals who have reentered society
and are still under criminal supervision, as well as those finished with their sentences. Faced with a law that covered both sex offender registrants and sex offender supervisees, Justice Kennedy does not distinguish between the two. Moreover, his emphasis on “convicted criminals” as particularly well-suited to benefit
from social media is a mirror image of the reintegrative and rehabilitative purposes
of supervision. As discussed, for instance, the Supreme Court has identified rehabilitation as one of the “primary goals” of probation.156 Though Justice Kennedy
does not speak explicitly to First Amendment rights of supervisees, the premise of
his opinion—social media is a particularly powerful channel of First Amendment
expression and the right to receive information for those reintegrating into
151. Brief for Respondent in Opposition to Certiorari at 25, Packingham, 137 S. Ct. 1730 (No.
15-1194).
152. Brief for Petitioner at 27, Packingham, 137 S. Ct. 1730 (No. 15-1194) (“[W]hatever restrictions are permissible when subjecting a person to government supervision pursuant to a lawful
sentence, petitioner and other registrants, who are ‘no longer on the “continuum” of state-imposed
punishments’… are entitled to ‘the full protection of the First Amendment.’”) (quoting Doe v. Harris,
772 F.3d 563, 570, 572 (9th Cir. 2014) (citing Samson v. California, 547 U.S. 843, 848 (2006))).
153. Id. at 52. The brief also references the rights retained by individuals while they are incarcerated. Id. at 27–28 (citing Pell v. Procunier, 417 U.S. 817, 822 (1974)).
154. Packingham, 137 S. Ct. at 1735 (emphasis added).
155. Id. at 1737.
156. United States v. Knights, 534 U.S. 112, 119 (2001). See supra note 50 (discussing the
goals of probation). See also Morrissey v. Brewer, 408 U.S. 471, 495 (1972) (“Under modern concepts of penology, paroling prisoners is part of the rehabilitative aim of the correctional philosophy.”).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

695

society—is plainly applicable in the supervision realm.
Neither of the two instances where Justice Kennedy notes that the North Carolina law covers individuals free from penal supervision restricts the opinion’s
broad coverage. First, Justice Kennedy writes in parentheses, “Of importance, the
troubling fact that the law imposes severe restrictions on persons who already have
served their sentence and are no longer subject to the supervision of the criminal
justice system is also not an issue before the Court.”157 Second, a few sentences
later, he writes: “It is unsettling to suggest that only a limited set of websites can
be used even by persons who have completed their sentences.”158 Neither of these
statements explicitly cabin the holding to a particular class of people, an option
clearly available to the Court.159 The first statement is somewhat cryptic, but its
words have a plain meaning: the Court is not ruling on whether the statute’s applicability to individuals finished with their sentences makes a constitutional difference. Indeed, in deciding that Packingham was applicable to a challenge
brought by sex offenders against Internet monitoring conditions, one district court
in Colorado cited this statement from Packingham as evidence that, as a matter of
first impression, it would need to answer this question the Supreme Court had left
open.160 Nor does the second reference cordon off the holding to sex offender
registrants exclusively. Indeed, in some individual cases, it could make sense to
limit the set of websites certain individuals under supervision may visit, while
such a restriction would be inappropriate for someone technically not under supervision.
Yet the next reason for not waving Packingham away in the supervision context is precisely because technical status does not tell the whole story here. That
is, beyond the textual signs that Packingham covers a broader class of individuals,
as a descriptive matter, offenders who have “completed” their sentences and supervisees who are still subject to restrictive conditions are treated similarly in the
criminal legal system. Even if Packingham is understood formally to apply only
to sex offender registrants, the similar impingements on liberty in the registrant
and supervisee contexts should extend Packingham’s logic to the latter group as
well. Without question, the law classifies these two groups differently: only supervisees are still serving out a criminal sentence, as supervision conditions are
considered part of the sentence, not a substitute for it.161 Yet as a practical matter,
both groups live under constant supervision and the threat of imprisonment. Sex
offender registrants nationwide are subject to criminal penalties for failure to
157. Packingham, 137 S. Ct. at 1737.
158. Id.
159. See Doe v. Prosecutor of Marion Cty., 705 F.3d 694, 703 (7th Cir. 2013) (striking down
a statute barring sex offenders from accessing the Internet, but clarifying that “this opinion should
not be read to affect district courts’ latitude in fashioning terms of supervised release”).
160. See Millard v. Rankin, 265 F. Supp. 3d 1211, 1228 (D. Colo. 2017).
161. See supra Part I.

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

696

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

comply with SORNA’s burdensome requirements. Under SORNA, any individual
defined by state or federal law as a “sex offender” who “knowingly fails to register” may be “fined under [the statute] or imprisoned not more than 10 years, or
both.”162
The rejoinder to this is doctrinal: only parolees and probationers have been
formally denied full constitutional rights, while sex offender registrants have
not.163 But reality belies this technical difference in who possesses more rights.
As a matter of public policy, extraordinarily burdensome restraints on sex offender
registrants’ liberty have repeatedly been deemed nonpunitive and upheld.164 And
in many cases, state versions of SORNA require individuals not under supervision
to report information to a probation officer.165 As three justices on the Supreme
Court have recognized, “The registration and reporting duties imposed on convicted sex offenders are comparable to the duties imposed on other convicted
criminals during periods of supervised release or parole.”166
Packingham does not alter this legal landscape of allowing far-reaching civil
liberties restrictions on sex offender registrants; it just says that these restrictions
should be narrowly tailored.167 So it should be with sex offender supervisees. This
is not to dispute the meaningful difference between a) being charged and convicted
with a new substantive offense as a registrant and b) having one’s supervision
revoked as a supervisee, nor to approve of the Supreme Court’s tacit treatment of
registrants as if they were still serving terms of criminal supervision.168 Rather, it
is to recognize that presently, these categories have much in common: both are
statuses requiring compliance with liberty-restricting conditions, under threat of
imprisonment.169 Furthermore, many supervisees, like registrants, are free from
162. See 18 U.S.C. § 2250(a).
163. See Morrissey v. Brewer, 408 U.S. 471, 477 (1972).
164. See, e.g., Smith v. Doe, 538 U.S. 84, 105–06 (2003). For more on constitutional problems
in SORNA, see Corey Rayburn Yung, One of These Laws Is Not Like the Others: Why the Federal
Sex Offender Registration and Notification Act Raises New Constitutional Questions, 46 HARV. J.
ON LEGIS. 369, 370 (2009) (noting that “federal courts across the nation have rubber stamped
SORNA’s provisions,” including the penalty of ten-year imprisonment for failure to register).
165. See Doe v. Kentucky ex rel. Tilley, 283 F. Supp. 3d 608, 610 (E.D. Ky. 2017) (noting
that Kentucky’s Sex Offender Registration Act required the petitioner, an individual previously convicted of a sex offense who was not under criminal supervision, to report his Internet identifiers to a
probation office).
166. Smith, 538 U.S. 84, 111 (2003) (Stevens, J., dissenting in part and concurring in part).
Justice Stevens wrote alone, but Justices Ginsburg and Breyer, dissenting separately, made an identical point: “[R]egistration and reporting provisions [for sex offender registrants] are comparable to
conditions of supervised release or parole.” Id. at 115 (Ginsburg, J., dissenting).
167. Contra Goldberg & Zhang, supra note 83, at 59, 86 (“[T]he Court recognized . . . [that]
registrants who have finished their sentence stand on the same footing as individuals who have exited
the criminal justice system after convictions for nonreportable offenses or those of us who have no
criminal justice history.”).
168. See supra note 166.
169. A functional, rather than formalistic, assessment of Packingham suggests that restrictions

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

697

punitive conditions. Recall the discussion in Part I on the federal supervision system’s focus on rehabilitation, deterrence, and incapacitation, excluding retribution
as a purpose of supervised release.170 While those on supervised release may still
be serving out their criminal sentences, they are no longer subject to punitive terms
of incarceration, just as registrants are typically free from punitive measures being
imposed on them.171 As a general matter, both groups can only be subjected to
narrowly tailored conditions that are designed to further underlying, nonpunitive
governmental interests. If registrants and supervisees are subject to similar restrictions on their liberty, then even if Packingham’s narrow tailoring holding is
nominally addressed to the former group, it should apply to the latter as well.
At this point, a pedant might say: That first, textual argument confuses holding for dicta, and the second, practical argument is aspirational, not doctrinal—
no court has ever found supervisees to be on the same constitutional footing as
registrants. Doctrinally, the pedant says, Packingham only held that a criminal
statute banning social media fails narrow tailoring. But the pedant’s argument ignores established doctrine on judicial scrutiny of supervision conditions, which
says that supervisees, like registrants, enjoy the protection of narrow tailoring
principles. Narrow tailoring, whether required by statute, such as 18 U.S.C. § 3583
for federal supervised release conditions, or a tier of First Amendment scrutiny, is
a common requirement of supervision conditions.172 There is no federal circuit in
which supervision conditions are reviewed with “rational basis”-level deference;
on liberty should be narrowly tailored, regardless of whether the individual is on supervision or
incarcerated. Although this Article does not take up the precise question of prisoners’ Internet rights
after Packingham, well-established precedent guarantees First Amendment protections for prisoners.
See Pell v. Procunier, 417 U.S. 817, 822 (1974) (“[A] prison inmate retains those First Amendment
rights that are not inconsistent with his status as a prisoner or with the legitimate penological objectives of the corrections system.”); Turner v. Safley, 482 U.S. 78, 84 (1987) (“Prison walls do not
form a barrier separating prison inmates from the protections of the Constitution.”).
170. See supra Part I.
171. Ignoring this similarity is another way for courts to distinguish Packingham from the
supervision context. For instance, in ruling that Packingham did not apply to supervision condition
challenges, an Eleventh Circuit panel recently noted:
[U]nlike the condition imposed on [an individual on federal supervised release]
for his past behavior, the statute at issue in Packingham was prospective: rather
than simply punishing a past crime, the statute there made it a new felony for a
person to use all social-media outlets, even though that person had had all impingements upon his constitutional rights lifted by fully serving the prior sentence.
United States v. Antczak, 753 F. App’x 705, 715 (11th Cir. 2018). This begs the question: in
what circumstance would it be permissible, in the panel’s view, for a social media ban to “simply
punish[] a past crime”? The sentence structure in this quotation suggests that the Circuit may (incorrectly) view supervision conditions, but not registration restrictions, as incorporating a punitive rationale. As discussed above and in Part I, this is contrary to the stated purpose of supervised release.
See 18 U.S.C § 3583(d)(1).
172. For further discussion of this requirement in supervision conditions, see supra Part II.A.

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

698

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

narrow tailoring, as shown above, is uniformly a part of the analysis. In other
words, supervisees do not need Packingham to demonstrate that social media restrictions imposed on them must be narrowly tailored. Where Packingham is relevant is on a point it makes more generally: social media bans are not narrowly
tailored to a significant governmental interest. And there is no claiming that the
principle of narrow tailoring does not apply in the supervision context—decades
of case law say otherwise.
C. How Does Packingham Strengthen Parole and Probation Challenges?
Assuming Packingham bears on supervision cases, whether as a directly controlling decision or as persuasive authority, what work does the opinion do for
those subjected to social media restrictions as a condition of their supervision?
It would be reasonable to think of Packingham as having minimal impact on
future cases dealing with social media restrictions, despite its flowing rhetoric.173
Sweeping language does not necessarily mean sweeping change to First Amendment doctrine; it may be fitting language for invalidating an egregiously unconstitutional law, or it may simply be Justice Kennedy writing like Justice Kennedy.174
As support for this view, the opinion does not make any apparent changes to how
courts choose intermediate or strict scrutiny, a point to which Justice Alito alludes
in his concurrence.175 After extensive briefing from the parties over whether the
statute was content-neutral, content-based, or something in-between,176 the Court
did not decide whether the North Carolina statute was content-neutral; it simply
held that “[e]ven making the assumption that the statute is content-neutral and thus
subject to intermediate scrutiny, the provision cannot stand.”177 It is a short
173. This is different from the question of its impact on future cases dealing with cyberspace
as a public forum, see supra note 17.
174. See Erin Daly, The New Liberty, 11 WIDENER L. REV. 221, 246 (2005) (noting Justice
Kennedy’s “grandiose” writing style); see generally Panel 2: Justice Kennedy’s Prose-Style and
Substance, 35 Ga. St. U. L. Rev. 907, 933 (2019) (discussing the intersection of Justice Kennedy’s
“grand style” with his substantive judicial philosophy).
175. Packingham v. North Carolina, 137 S. Ct. 1730, 1743 (2017) (Alito, J., concurring) (“After noting that ‘a street or a park is a quintessential forum for the exercise of First Amendment rights,’
the Court states that ‘cyberspace’ and ‘social media in particular’ are now ‘the most important places
(in a spatial sense) for the exchange of views.’ The Court declines to explain what this means with
respect to free speech law . . . .”).
176. See Brief for Petitioner at 40, Packingham, 137 S. Ct. 1730 (No. 15-1194) (“This Court
does not treat measures like Section 202.5 that disfavor the speech of a subset of speakers as contentneutral time, place, or manner restriction.”); Brief of Respondent at 17, Packingham, 137 S. Ct. 1730
(No. 15-1194) (“Section 202.5’s operation does not turn on ‘the topic discussed or the idea or message expressed.”).
177. Packingham, 137 S. Ct. at 1736. While the Court declined to decide whether the law is
content-neutral, at least one early lower court decision in Packingham’s wake seems to interpret it,
perhaps mistakenly, as holding otherwise. See Sandvig v. Sessions, 315 F. Supp. 3d 1, 12 (D.D.C.
2018) (“The Packingham Court . . . employ[ed] intermediate scrutiny because the law was contentneutral.”).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

699

opinion tackling no difficult questions—as one commentary puts it, “an easy
case.”178 And the holding—that a social media ban for a class of sex offenders is
an overbroad restriction of speech—seems to affirm, but not add onto, the analysis
in Part II. Yet the opinion strengthens arguments against supervision conditions in
several tangible ways.179
1.

Packingham Says That Social Media Bans Are a Per Se Impingement
on First Amendment Rights.

First, perhaps more important than identifying and then condemning the obvious breadth of the North Carolina statute,180 the Court clearly identifies the statute as covering speech, and then condemns the law as a speech deprivation.
A bit of background on why this matters, and why the Court was not merely
reciting a truism, is in order. To prove that a deprivation is too extreme, one must
show that a deprivation exists at all. In the case of Internet restrictions, it must be
shown that taking away an individual’s Internet access is objectively a severe deprivation of liberty. Perhaps surprisingly, Packingham is the first case to state this
plainly, recognizing that access to the Internet is a fundamental means of First
Amendment expression and access to information, regardless of individual circumstance.
Prior to Packingham, no Supreme Court decision—including Reno v. American Civil Liberties Union—acknowledged the centrality of the Internet to speaking freely and accessing information, two central First Amendment tenets. To be
sure, Reno, the Court’s first statement on the Internet, exactly 20 years before
Packingham, did marvel at this “international network of interconnected computers.”181 Yet Reno, and other early decisions discussing the Internet, focused more
on the diversity of what one could do and express on the Internet than on its centrality to expression.182 After all, when Reno was decided in 1997, the Internet
was still considered a “wholly new medium of worldwide human communication.”183 Several decisions in subsequent years noted the rapid expansion of
178. Goldberg & Zhang, supra note 83, at 64.
179. The Court does briefly gesture toward the “wrong people” argument, see supra Part II.B.,
explicitly noting a mismatch problem near the beginning of the opinion: “At no point during trial or
sentencing did the State allege that petitioner contacted a minor—or committed any other illicit act—
on the Internet.” Packingham, 137 S. Ct. at 1734.
180. Packingham, 137 S. Ct. at 1737 (“[T]he statute here enacts a prohibition unprecedented
in the scope of First Amendment speech it burdens.”).
181. 521 U.S. 844, 849 (1997).
182. See id. at 852 (“[T]he content on the Internet is as diverse as human thought.”); Ashcroft
v. Am. Civil Liberties Union, 535 U.S. 564, 566 (2002) (“While ‘surfing’ the World Wide Web, the
primary method of remote information retrieval on the Internet today, individuals can access material
about topics ranging from aardvarks to Zoroastrianism.”) (citation omitted).
183. Compare Reno, 521 U.S. at 850 (noting that in 1996 that the Internet had approximately
9.4 million hosts and 40 million users), with Brief for Electronic Frontier Foundation et al. as Amici

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

700

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

Internet usage in the United States.184 But it was not until 2017 that the Supreme
Court directly revisited the relationship it had begun examining in Reno between
First Amendment-protected activity and the Internet.
When Reno was decided, this international network of interconnected computers was still in its infancy. Much changed in the “twenty year hiatus” between
that decision and Packingham.185 The latter, for all its soaring language about the
Internet, includes no far-reaching examples demonstrating the Internet’s expansive range. There are no starry-eyed references to the Internet’s ability to yield
information on “topics ranging from aardvarks to Zoroastrianism,” as there were
in Reno.186 Instead, the Packingham opinion focuses on how integral social media
has become—“Seven in ten American adults use at least one Internet social networking service”—for the most important social functions: “debat[ing] . . . politics,” “look[ing] for work,” and “petition[ing] . . . elected representatives.”187 The
Court goes so far as to name the Internet, and “social media in particular,” as “the
most important places (in a spatial sense) for the exchange of views.”188
Why is a statement acknowledging the Internet’s centrality in society important? What difference does it make to have the Supreme Court’s imprimatur
on what we all already know to be true? Is it not obvious that taking away one’s
Internet access poses an extreme deprivation to First Amendment rights? Apparently, it is not. Several lower court cases prior to Packingham had implied, and in
some cases explicitly stated, that restricting Internet access did not always constitute a severe deprivation of a supervisee’s liberty. Numerous circuits weighed supervisees’ lack of demonstrated need for the Internet against them in deciding the
permissibility of the broad Internet restrictions as conditions of supervision. For
example, in United States v. Angle, the Seventh Circuit considered whether a supervised condition barring “personal access to computer Internet services” from a
defendant convicted of multiple sex offenses was appropriate.189 The court justified upholding the restriction in part due to aspects of the defendant’s work history: “[H]is use of the Internet was not integrally connected to his profession as
he was previously employed as a salesman and mechanic.”190 Mechanics don’t
Curiae in Support of Petitioner at 5, Packingham, 137 S. Ct. 1730 (2017) (No. 15-1194) (noting in
2017 that the Internet has over 1 billion hosts and over 3.5 billion users).
184. See United States v. Am. Library Ass’n, 539 U.S. 194, 199 (2003) (“By 2000, 95% of
the Nation’s libraries provided public Internet access.”); Ashcroft v. ACLU, 542 U.S. 656, 671
(2004) (noting the dramatic changes in Internet technology between the district court holding from
1999 and 2004).
185. First Amendment-Freedom of Speech-Public Forum Doctrine-Packingham v. North Carolina, supra note 14, at 233.
186. Ashcroft, 535 U.S. at 566.
187. See Packingham v. North Carolina, 137 S. Ct. 1730, 1735 (2017).
188. Id. at 1743.
189. 598 F.3d 352, 360 (7th Cir. 2010).
190. Id. at 361.

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

701

need the Internet, the logic goes, so a former mechanic barred from it can’t complain of foregone employment prospects. The Seventh Circuit is not alone in
weighing blue collar work history against a supervisee-defendant who claims a
need for the Internet.191
Which is the most striking aspect of this reasoning? Its implication that the
defendant’s history will determine the boundaries of his future rehabilitation? Its
devaluation or omission of all the ways the defendant would need or benefit from
the Internet outside of work?192 Its faulty assumption that blue collar jobs or workers do not require the Internet? Its implication that defendants with professional
interests in using the Internet who have been convicted of sex offenses are more
justified in regaining Internet access? The Packingham Court does not provide an
answer. But it does deem anachronistic even asking the question, ‘does this person
really need the Internet?’ In the 20 years since Reno, the Internet has become impervious to “mile wide, inch deep” taunts. Its ever-growing centrality makes it
increasingly difficult to justify broad restrictions that kick people offline.
Some courts began to realize this change before Packingham. In a 2010 case
involving a broad Internet restriction, the D.C. Court of Appeals addressed United
States v. Paul, a 2001 decision where the Fifth Circuit upheld an absolute ban on
Internet access. In the intervening period since 2001, the D.C. Circuit held, “the
computer and internet have permeated everyday life in ways that make a restriction on their use far more burdensome than when Paul was decided.”193
Packingham enshrines this point in Supreme Court precedent. “These websites can provide perhaps the most powerful mechanisms available to a private
citizen to make his or her voice heard,” the opinion reads.194 With this focus on
political or civic speech, the Court defends social media access generally, without
regard for specific circumstances making it more or less likely the restricted individual will actually use the Internet. Listening to and engaging in political speech
does not require an extensive background in political speech—it only requires a

191. See also United States v. Granger, 117 F. App’x 247, 249 (4th Cir. 2004) (“[T]he great
majority of [his] work history involves manual labor; he has held jobs as a pipe fitter, driller, and
field hand, with only very brief interludes as a cashier or clerk. . . . [His] ability to return to similar
gainful employment would not be greatly hampered by this special condition.”); United States v.
Alvarez, 478 F.3d 864, 868 (8th Cir. 2007) (“Finally, unlike the computer consultant defendant in
U.S. v. Mark, 425 F.3d 505, 509 (8th Cir. 2005), Alvarez’s employment history (which included
work as a stocker at a Target store and sporadic employment at Wal–Mart and fast food establishments) does not indicate that he has a particular day-to-day vocational need for Internet access.”);
United States v. McDermott, 133 F. App’x 952, 954 (5th Cir. 2005) (“McDermott also asserts that,
due to his employment history, a computer and Internet access will be essential to his ability to earn
a living. This argument is speculative at best given that McDermott is now 62 years old.”).
192. In fairness, several cases inquiring into a supervisee’s need for the Internet on the job ask
this as one of many questions regarding the supervisee’s need for the Internet.
193. United States v. Russell, 600 F.3d 631, 638 (D.C. Cir. 2010).
194. Packingham v. North Carolina, 137 S. Ct. 1730, 1737 (2017).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

702

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

8/19/19 2:25 PM

[Vol. 43:663

Twitter handle.195
Relatedly, the Court also spoke—albeit indirectly—to the “ample alternative
means” requirement of First Amendment scrutiny. While omitting reference to
Ladue,196 the Court stated that social media offers “relatively unlimited, low-cost
capacity for communications of all kinds,” and touted its modern centrality in no
uncertain terms.197 The discussion above applies here with equal force: the Packingham Court regards social media as an objectively important tool that “can provide perhaps the most powerful mechanisms available to a private citizen to make
his or her voice heard.”198 In its most emphatic nod to Ladue, the Court concludes
its narrow tailoring analysis by remarking: “In sum, to foreclose access to social
media altogether is to prevent the user from engaging in the legitimate exercise of
First Amendment rights.”199 Thus, to the extent the “ample alternative channels”
question is tied to future challenges on social media restrictions, the Packingham
opinion is clear that social media has no adequate alternative.
The categorical value of social media access does not mean courts are suddenly precluded from imposing restrictions on it. Packingham simply clarifies that
in determining whether a social media restriction is too great a deprivation of liberty, there can be no debate over whether such a deprivation is severe. In other
words, Packingham creates a floor. Perhaps some defendants will have additional
reasons for why depriving them of social media access would be especially harmful, but every defendant will, at minimum, be able to argue persuasively that this
deprivation would impinge on crucial First Amendment freedoms.200
195. “Handle” is the colloquial term for a username on Twitter. See Leslie Walker, “Twitter
Language: Twitter Slang and Key Terms Explained,” Lifewire.com, https://www.lifewire.com/twitter-slang-and-key-terms-explained-2655399 [https://perma.cc/N8PX-SUB7] (last visited April 27,
2019).
196. Mr. Packingham’s counsel, and amici supporting him, had asked the Court to clarify the
boundaries of Ladue and to find that the North Carolina law had foreclosed a “means of communication that is both unique and important.” See Brief Amici Curiae of Prof. Ashutosh Bhagwat et al.
in Support of Petitioners on Petition for Writ of Certiorari to the North Carolina Supreme Court at
12, Packingham, 137 S. Ct. 1730 (2017) (No. 15-1194) (seeking Supreme Court review on how the
“ample alternative channels” requirement should be understood); Brief for Petitioner, Packingham,
No. 15-1194, at 54–56. In avoiding citation to Ladue, the Supreme Court maintained Ladue’s long
drought: as of 2019, it has never again been referenced by the Court to impose an adequate alternatives requirement.
197. Packingham, 137 S. Ct. at 1735 (quoting Reno v. Am. Civil Liberties Union, 521 U.S.
844, 870 (1997)).
198. Id. at 1737.
199. Id.
200. By contrast, Goldberg & Zhang too easily divorce the Packingham opinion’s Internet
musings from its embrace of the proposition that registrants are entitled to full constitutional rights.
See Goldberg & Zhang, supra note 83, at 84 (stating, in response to the Court’s language on the
centrality of the Internet to First Amendment-protected activity: “But none of this broke new
ground.”). The ground may have already cracked in Reno, but Packingham broke it open. The principle of the “free-speech equality” idea that Goldberg & Zhang find permeating throughout the Packingham opinion is due, in part, to a revolution in how fundamentally we use the Internet, a revolution

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

703

2. Packingham Suggests That “Social Media” is Too Broad a Category
Somewhere on the line between a vagueness and an overbreadth argument,
the Packingham opinion cautions against broadly proscribing “social media” without further definition. A main source of disagreement between the Packingham
majority and concurrence was over the need for precision in defining this term.
Justice Kennedy’s opinion states that the Court “need not decide the precise scope
of the statute,” as it is “enough to assume that the law applies . . . to social networking sites ‘as commonly understood’—that is, websites like Facebook,
LinkedIn, and Twitter.”201 But Justice Kennedy prefaces this discussion with an
entire section devoted to the First Amendment importance of “social media in particular,” unmoored from the North Carolina statutory language.202 Thus, the decision becomes less about North Carolina’s four-part definition of a “commercial
social networking Web site,” and more about the impropriety of banning access to
“social media.”
Unsurprisingly, Justice Alito is quick to assure readers that Justice Kennedy’s
broad proclamations about social media are “undisciplined dicta.”203 Dicta or not,
they concern him: “[T]his language is bound to be interpreted by some to mean
that the States are largely powerless to restrict even the most dangerous sexual
predators from visiting any internet sites.”204 Rather than blessing the undefined
expanses of social media with First Amendment holy water, Justice Alito sticks to
the law in front of him, which “covers websites that are ill suited for use in stalking
or abusing children,” such as Amazon.com, the Washington Post’s website, and
WebMD.205 Facebook, the primary source at issue in this case, does not appear in
Justice Alito’s list, a sign of how the warier justices might treat a no-Facebook
provision.206 Per the majority, however, conditions that, without further definition, restrict access to “social media” will be deemed to sweep in an array of protected speech and could be struck down.
* **
Existing case law could have itself supported a strong challenge to these types
of mismatched restrictions. Yet as Mr. Packingham’s counsel wrote of social media bans, “These laws have provoked barely a judicial peep.”207 If the law is on
their side, why don’t individuals subject to mandatory restrictions seek their invalidation? And why do burdensome, mismatched discretionary conditions persist? There are several reasons for the dearth in case law on this subject. On the
which had not taken place at the time of Reno’s writing. Id. at 87.
201. Packingham, 137 S. Ct. at 1737.
202. Id. at 1735.
203. Id. at 1738 (Alito, J., concurring).
204. Id.
205. Id. at 1743 (Alito, J., concurring).
206. See id. at 1741–43 (Alito, J., concurring).
207. See Goldberg & Zhang, supra note 83, at 60.

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

704

8/19/19 2:25 PM

N.Y.U. REVIEW OF LAW & SOCIAL CHANGE

[Vol. 43:663

subject of mandatory restrictions, these may be more resistant to challenge than
as-applied challenges to discretionary conditions. While narrow tailoring problems are evident in the state statutes discussed in Part II.B, challenging a discretionary condition may pose a lesser burden to a burdened supervisee who only
possesses knowledge of his individual restriction. Moreover, it is possible that
judges and supervisory officers in these states with mandatory restrictions do not
enforce the laws strictly, and instead discretionarily exempt certain individuals
under their supervision from broad social media restrictions. Lastly, with regard
to challenges against both mandatory and discretionary conditions, the space for
such a challenge may have been unclear before Packingham. This uncertainty
should change in the wake of the Supreme Court’s decision. With a guarantee of
narrow tailoring to assess a social media ban, a recognition of speech deprived,
and harsh words for a broad, vague ban on “social media,” Packingham could shift
the momentum.
IV.
CONCLUSION
Parole and probation were once considered acts of governmental grace. Any
individual autonomy under this status was a gift, not an entitlement. With the rise
of statutory supervised release and rehabilitative parole, this doctrine supposedly
died. But work remains to bury it—and signs are not pointing toward legislative
fixes. Amid shifting attitudes around other areas of criminal law and incarceration,
social attitudes toward individuals convicted of sex offenses remain generally
fearful and retributive. In this context, legislative protections from overly restrictive supervision conditions for these individuals are unlikely to appear in the near
future. Indeed, even the most inspiring of recent democratic steps to restore rights
to those with criminal records have excluded those whose convictions were for
sex offenses.208 In the meantime, legal advocates can marshal new precedent like
Packingham on a variety of levels to fight conditions that too greatly impinge on
civil liberties. Public defenders can file motions objecting to Internet bans placed
on their clients as conditions of parole and probation, and can appeal incarceratory
sentences which include overbroad conditions for post-conviction supervised release.209 Civil liberties organizations and law school clinics can bring facial

208. See Paul Wright, The Case Against Amendment 4 on Felon Voting Rights, TALLAHASSEE
DEMOCRAT (Sept. 19, 2018), https://www.tallahassee.com/story/opinion/2018/09/19/case-againstamendment-4-felon-voting-rights-opinion/1351689002/ [https://perma.cc/UQG6-3HFG ] (noting
that the movement to reverse felon disenfranchisement in Florida excluded individuals with murder
or felony sex offense convictions).
209. See, e.g., United States v. Carson, No. 17-3589, 2019 WL 2063371, at *3 (8th Cir. May
10, 2019) (defendant-appellant appealed—albeit unsuccessfully—sentence of 20 years of imprisonment followed by life term of supervised release, which included a prior approval restriction on
Internet access).

5 HUTT_43.4_V2.DOCX (DO NOT DELETE)

2019]

8/19/19 2:25 PM

OFFLINE

705

challenges to mandatory legislative bans or assist defenders’ efforts through amicus briefing.210 Public interest law firms can work with community advocacy
groups to build class action lawsuits against city, county, or state policies in imposing Internet restrictions.211 No matter what form the fight takes, Reverend
Sherman Manning, and others like him, need not wait for newfound social consensus to enjoy the rights to which they are entitled.

210. For example, the American Civil Liberties Union of New Jersey and the Rutgers Constitutional Rights Clinic Center for Law & Justice supported as amici an as-applied due process challenge against an Internet ban as a condition of parole, which resulted in the Supreme Court of New
Jersey striking down the condition. See J.I. v. New Jersey State Parole Bd., 228 N.J. 204, 212 (2017).
211. This model was recently successful in Illinois, where a class composed of “adult criminal
sex offenders who have completed their terms of imprisonment and attempted to comply with the
requirements imposed upon them in securing a [residential] host site” prevailed on their motion for
summary judgment in challenging residency restrictions. Murphy v. Raoul, No. 16-C-11471, 2019
WL 1437880, at *17 (N.D. Ill. Mar. 31, 2019). The class, represented by Chicago-based civil rights
lawyers Adele Nicholas and Mark Weinberg, alleged in part that the Illinois Department of Corrections’ practice of forbidding sex offender parolees from transitioning to homes with internet access
misuses the Department’s statutory authority. Class Action Complaint, Murphy v. Madigan, No. 16C-11471 at ¶ 153 (N.D. Ill. Dec. 19, 2016). Nicholas and Weinberg are “partner attorneys” of Illinois
Voices, a volunteer-led, non-profit organization which “promotes the elimination of sexual abuse
and the preservation of civil rights” for individuals convicted of sex offenses, cooperating with local
attorneys to bring legal challenges in Illinois. Mission and Values, ILLINOIS VOICES,
http://www.ilvoices.org/mission-statement.html [https://perma.cc/CY5B-HEYL] (last visited May
25,
2019);
Legal
Efforts,
ILLINOIS
VOICES,
http://www.ilvoices.org/legal.html
[https://perma.cc/JR32-6XS3] (last visited May 25, 2019); About Illinois Voices, ILLINOIS VOICES,
http://www.ilvoices.org/about-us.html [https://perma.cc/N2C7-C2QZ] (last visited May 25, 2019).