State Laws Restricting Social Media Use by Sex Offenders Are Failing in Court
[Post by Venkat Balasubramani with comments by Eric]
Statutory schemes in three different states intending to regulating the online activities of convicted sex offenders have meet with judicial disapproval.
Doe v. Nebraska, 09CV456 (D. Neb. Oct. 17, 2012):
Nebraska’s statute (1) required registrants to disclose their device identifiers and online profiles; (2) required registrants to consent to searches and the installation of monitoring apparatus and (3) criminalized use of sites that are accessible by minors.
Ban on internet use: Section 28.322.05 banned sex offenders from using any “social networking website, instant messaging, or chat room service” that also allows users who are less than 18 to access those services. Citing to an array of facts and figures regarding the ubiquity of social networks and utilities, the court says that the statute restricts the affected individuals from using “an enormous portion of the internet” to engage in expressive activity. The court notes that the ban is not dependent on any past use of utilities to commit crimes, or any particular online risk posed by the offenders. The ban is overly broad and violates the First Amendment.
The court also picks apart the definitions employed by the statute and says that the statutory definitions render it unconstitutionally vague. The government offered a slew of narrowing constructions to different terms in the statute, but the court says that these suggestions further illustrated the vagueness underlying the statute. For example, the statute banned “instant messaging” services—that were defined as services that enabled instantaneous text transmissions—but the State claimed that this did not include text messaging services. The court was charitable toward the legislature, but at a certain point the snark just broke through:
Without intending to be unkind, the fourth suggested construction is laughable. It states that ‘virtually instantaneous’—for purposes of ‘instant messaging’ services or ‘chat rooms’—means ‘real time.’ What in the world does ‘real time’ mean? Particularly when it comes to ‘text messages’ sent through ‘instant messaging’ services, the substitution of the words ‘real time’ for ‘virtually instantaneous’ is of no help whatever in clarifying the glaring ambiguity in the statute. The proposed construction is very much like a dog chasing its tail—the dog and the tail simply turn in a humorous circle.
The court says that, in addition to being overly broad, the ban does not leave open alternate channels of communication. While the State argued that sex offenders could still obtain information through news media, the court says that not being able to access information via Twitter and Facebook is as good as a ban—it doesn’t matter that you can access the news in the next day’s edition of the newspaper. Similarly, not being able to use Skype means not being able to videoconference with family members. According to the court, there’s no substitute.
The court rules this provision facially unconstitutional.
Disclosure of identifiers: The statute requires offenders to disclose device identifiers as well as “internet communication identifiers”. A related statutory provision also required offenders to consent to a search of his or her devices. With a cite to McIntyre and a nod to the fact that “anonymity is a shield from the tyranny of the majority,” the court says that these provisions also do not pass muster. The court finds particularly troubling the fact that an offender is required to disclose any blogs or sites maintained by the person – i.e., the disclosure provisions effectively eviscerate an offender’s right to engage in anonymous online speech. And the court is clear that even registered sex offenders have this right.
The court also rules this provision facially unconstitutional.
Ex Post Facto Challenge: The court also says that the statute violates the ex post facto clause. The key question is whether the statute has a punitive intent or merely a civil regulatory scheme. Citing to statements from the sponsoring lawmaker (and the fact that the State elected to exercise legislative privilege to block inquiry into the purpose of the bill), the court says the statute is punitive in nature and violates ex post facto with respect to people who had been convicted prior to the effective date.
The court does dismiss as unripe the Fourth Amendment challenge brought against the consent to search provisions by those who were not on probation. (In any event, the court ruled these provisions unconstitutional on other grounds.)
Doe v. Harris, C12-5713 (N.D. Cal. Jan. 11, 2013):
California voters overwhelmingly approved the “Californians Against Sexual Exploitation Act” (Proposition 35). The statute requires sex offenders who fall under California’s sex offender registration program to provide the following:
- a list of all “internet identifiers established or used” by the person
- a list of all “service providers used by the person”
Internet identifier is further defined as:
an electronic mail address, user name, screen name, or similar identifier used for the purpose of Internet forum discussions, Internet chat room discussions, instant messaging, social networking, or similar communication.
The statute also required registrants to provide (in writing) any changes to the above information. The information is to be provided to the agencies that otherwise deal with the registrant, but those agencies would make the information available to the Department of Justice. The court initially entered a TRO, but subsequently issued a preliminary injunction, enjoining enforcement of the statute pending outcome of the litigation.
Level of scrutiny: First, the court says that the statute is not subject to strict scrutiny because it is content neutral, even though it affects a class of speakers.
Construing the statute: Second, the court narrows the statute in accordance with the government’s concessions: (1) the only service providers a registrant needs to report are those with whom the registrant actually has an account (e.g., if you are at the library and Time Warner is the ISP that allows you to access the internet from the library, you need not report this); (2) only identifiers that are actually used to post content or communicate are required to be reported (e.g., your Amazon account information that you use to purchase books is not required to be reported, assuming you also do not use that profile to comment or interact online).
Is the statute “narrowly tailored”?: The court says that it’s conceivable that the registration requirement generally advances the stated government interest of preventing future crimes by registrants. The court notes the hypothetical example of being able to cross-check a person who uses an online identity to recruit against the database of registrants. However, the court says that narrow tailoring in this context also requires restrictions on when law enforcement can access the information and what they can do with it. Here the law says that law enforcement can disclose identify information “when necessary to ensure public safety.” The law was similar to statutes in other states (Georgia and Utah) that have been struck down.
The court is also concerned that the statute will have a chilling effect—failure to report is a criminal violation, punishable by up to three years in prison. The statute also has the effect of depriving registrants of their online anonymity.
The court points to the fact that the state’s own assessment process classified a chunk of the offenders as having a “low to moderate” risk of re-offending. The government could not explain why these offenders should be treated the same as higher risk offenders. There was also no data of how likely offenders are to commit future offenses using the internet. Similarly, the court says that some data is available to the effect that online exploitation is much less likely to occur on sites that involve discussion of political or social issues. It’s unclear there’s adequate justification for requiring registration for these types of sites or services (even for those whom the data ostensibly says are likely to re-offend online).
In closing the court says:
The Case ACT provisions extend to [websites dedicated to discussion of public, political, and social issues], and registrants are likely to be chilled from engaging in legitimate public, political, and civil communications for fear of losing their anonymity. As a Nebraska district court forcefully stated, a requirement that sex offenders report to the government all communications on blogs and websites puts a stake through the heart of the First Amendment’s protection of anonymity [and] surely deters faint-hearted offenders from expressing themselves on matters of public concern.
Interestingly, the statute was sponsored by Facebook’s ex-Chief Privacy Officer.
Doe v. Indiana, 2013 WL 238735 (7th Cir. Jan. 23, 2013):
The Indiana statute in question prohibited sex offenders from:
knowingly or intentionally using a social networking web site or an instant messaging or chat room program that the offender knows allows a person who is less than eighteen (18) years of age to access of use the web site or program.
Doe sued on his own behalf and on behalf of a putative class of offenders who were required to register but were not on any form of supervised release. The district court found that the statute burdened more speech than necessary, but upheld the statute, finding that Doe didn’t “furnish the court with workable [alternatives].” The Seventh Circuit disagrees, and says the statute does not pass First Amendment muster.
According to the Seventh Circuit, the issue was that the problematic activity was only a subset of the overall expression that the statute regulated (something the Nebraska and California courts focused on also). It was aimed at improper online communications between registered offenders and minors, but this was “a minuscule subset of the universe of social network activity.” Moreover, Indiana already had statutes on the books that were targeted at the particular ills at issue – it’s already a crime in Indiana to solicit minors, or engage in inappropriate communications with children. Penalties are increased when you do so using a computer.
The court does step back and say that it should be careful to not impose too high a standard on state legislatures. Some amount of over-inclusiveness can be justified by “administrabilty concerns,” but the statute as currently crafted can’t be justified on this basis.
The State also argued that existing statutes serve a different purpose than the social networking ban at issue in the case. According to the State, existing laws aim to punish those who have committed the crime of solicitation, while the ban aims to “prevent and deter.” The court is not persuaded by this distinction, saying that all laws punish those who have committed conduct that is proscribed by the laws, and this punishment is what deters. The court is also not sold on the argument that the social networking ban would be more effective because it would prevent would-be offenders from being on social networks in the first place. The court says that to the extent someone would break the law and solicit a child, they would be just as likely to break the law banning them from using a social networking site in the first place.
The court closes with a nod to the legislature that it’s not foreclosing a more carefully crafted statute, or even one that may be better supported by legislative facts or studies. To the extent a class of individuals “whose presence on social media impels them to solicit children,” the legislature could presumably enact legislation targeting these individuals.
Finally, the court says that nothing in its opinion should be read to limit the latitude of district courts in fashioning terms of supervised release.
These three cases, as well as laws aimed at Backpage from Washington and Tennessee, are more than enough to mark a trend. Legislatures are rightly concerned with online sex offenses, but are taking the blunt instrument approach to regulating the online activities of sex offenders. Courts are consistently saying that the First Amendment requires a more finely crafted approach. As Eric has highlighted before, state legislatures do not seem particularly adept when it comes to regulation online. Definitional terms are a major stumbling block. The Nebraska case in particular highlights this; the court had a field day poking holes in legislative definitions of things like “chat rooms” and “instant messaging”.
Though the three statutes differed significantly in their approach, in all three cases the courts found that only a subset of sex offenders were likely to engage in problematic conduct online, and regulation of sex offenders as a whole without reference to whether they were likely to engage in problematic acts online sweeps in too much First Amendment protected expression – i.e., sex offenders use the internet for normal activity just like everyone else. To the extent legislatures try to regulate the online activities of those whose sex offenses included an online component, they would probably have a much easier time doing so. However, even as to these individuals, courts will be concerned with any over-inclusiveness of a ban. On a related note, trying to restrict registrants’ access to sites that also allow access by those under 18 is not workable.
[Eric's addendum: There is actually social science on this question that could inform policy-makers if they actually did their homework. For example, I call your attention to Lee, Austin F., et al. Predicting hands-on child sexual offenses among possessors of Internet child pornography, 18 Psychol. Pub. Pol'y & L. 644-672 (2012):
CP [Child pornography] offenders appear to comprise a subgroup of sex offenders characterized by taxonomic heterogeneity. As Seto has pointed out on multiple occasions, those apprehended with CP have a sexual interest, if not a sexual preference, for children, and, given prevailing DSM criteria, are frequently diagnosable as pedophiles. Indeed, this same point was noted in U.S. v. Swarm–Dr. Mills and Dr. Saleh correctly agreed with the Bureau of Prisons’ memorandum that states, quote: “Paraphilias, including pedophilia, range in severity from a condition in which the individual experiences deviant sexual fantasies and urges, but did not engage in any victim contact, to individuals who act on their urges and fantasies . ..” (p. 21). Paradoxically, this group of pedophiles, as noted, is at low risk to commit hands-on sexual assaults of children. Those CP offenders that do sexually assault children are distinguished by a much higher degree of antisociality compared to those that refrain from such crimes. Moreover, those CP offenders that sexually assault children typically present as lower in educational and vocational achievement than those for refrain from such crimes. We found in the present study, e.g., that 21% of the Internet-only offenders were professionals, compared with only 8% of the “dual” offenders. Witt (2010) commented that “Studies have found that child pornography offenders are generally more educated, more intelligent, and have more stable work and relationship histories than contact sex offenders” (p. 4). Generally, these findings are consistent with the hypothesis that increased social and vocational competence inhibit the expression of antisocial behavior in IO-only offenders. By contrast, one could readily hypothesize that traits associated with Antisocial Personality Disorder (APA, 2000), such as deceitfulness, manipulativeness, impulsivity, aggressiveness, disregard for others, and impaired social emotions (remorse, guilt, and empathy), more likely found among offline offenders, are disinhibitory to committing a battery offense. Babchishin et al. (2011) concluded similarly, noting the presumptive importance of inhibitors and self-control in differentiating between online and offline offenders.]
The cases also highlighted internet anonymity and made clear that just because you are a sex offender does not mean that you lose your First Amendment right to speak anonymously online. As Eric points out below, sex offenders aren’t a class of individuals who have the most political clout, so it’s nice to see judges acting as a meaningful check on the legislature in this arena.
Efforts to ban sex offenders from social media sites, or require them to report their credentials, are at the intersection of three memes that have consistently undermined rational policy-making:
1) Protect the kids online. “Protect the kids” makes for great political soundbites, but let me ask a serious question: Can you point me to *any* law that was enacted under the conceit of protecting kids online that has actually had that effect? Politicians pontificate greatly on the need to protect kids online, but I’m drawing a blank on when they’ve succeeded with that effort.
2) Deprive sex offenders of civil liberties. As pariahs in our society, sex offenders have no one fighting for their civil liberties in legislative processes, so policy-makers can reduce their civil liberties to sub-human levels without any pushback. Fortunately, the courts are standing up to these systematic legislative efforts to functionally eliminate convicted sex offenders from our society, but only at significant cost.
3) Social media exceptionalism. When a new and important Internet medium emerges, policy-makers feel like they have to be “cutting-edge” and fix the purported problems in the emerging medium. But as these and other numerous cases have made abundantly clear, it is absolutely impossible to define social media as a subset of the Internet, so the bans on using social media become the functional equivalent of Internet bans and therefore are clearly unconstitutional. Yet, despite the resoundingly clear message from the courts, statute-drafters keep making the same damn drafting mistakes.
I am particularly disgusted with California’s Proposition 35 for three reasons:
1) It was a bait-and-switch on the public. The proposition was sold as an anti-human trafficking law, but the online sex offender restrictions went well beyond that topic.
2) I am not aware of any social science linking sex offenders online with human trafficking. As far as I know, the legislation hardwired assumptions about sex offender behavior that had absolutely no grounding in fact. This wasn’t even junk science; it’s ascientific policy-making.
3) Even though California has a deficit of billions of dollars, we taxpayers spent our money defending this law–despite the fact it was obviously unconstitutional on its face. Yay for wasted government resources!
I believe California’s voter initiative process is irreparably broken, and seeing shitty and unconstitutional initiatives like this get 80%+ voter approval provides more evidence of that.
Overall, the fact that states keep making the same systematic errors provides more evidence that states are terrible “laboratories of experimentation.” Rather than replicating the best state laws after the laws empirically demonstrate their worth, states embrace and propagate bad memes quickly–and without any evidence of efficacy–when it comes to regulating the Internet, multiplying the costs to judicially remedy the legislative pandering. We desperately need a moratorium on state laws that regulate the Internet until policy-makers can figure out why they keep making the same mistakes and build processes to overcome the forces causing the systematic errors. (Of course, I’d rather just get state legislatures out of the Internet regulation business altogether, but a moratorium would be a good start).