Skip to Content

Free Expression

What ‘Woodhull’ Won’t Change: Five Years of Chilling Effects Under FOSTA

By CDT Intern Clare Mathias

During the debate leading up to the passage of the Fighting Online Sex Trafficking Act (FOSTA), advocates warned that the statute’s broad language would silence speech about sex and sex work, creating risks for expression and physical safety. Five years later, this concern has largely come true: sex workers face discrimination from platforms and greater physical danger, and it’s significantly harder for everyone to talk about sex online. In July, a DC Circuit Court tried to remedy some of these harms by adopting a relatively narrow interpretation of FOSTA’s language. But even with this change, it will be nearly impossible to reverse the five years of chilling effects that FOSTA created. 

Woodhull v. US Decision 

Since FOSTA was introduced over five years ago, sex workers, civil liberties advocates (including CDT), and legislators have voiced concerns about the broad language in FOSTA prohibiting any online platform from “promoting” or “facilitating” sex trafficking. The law was not clear about what exactly fell into this novel standard, which drove many online services to heavily restrict all content about sex work. 

In July 2023, the DC Circuit Court of Appeals rejected the arguments from CDT and the parties in the case that FOSTA was unconstitutional because it jeopardizes constitutionally protected speech about sex and sexuality. Instead, the court ruled that “promote and facilitate” must be understood narrowly as equivalent to the more common “aiding and abetting” standard of liability. 

By interpreting “promote and facilitate” in this way, the court set a high bar for prosecution under FOSTA. As the Supreme Court explained in Twitter v. Taamneh, aiding and abetting requires “conscious, voluntary, and culpable participation in another’s wrongdoing”—a prosecutor must prove that the platforms not only knew about the illegal activity but that they themselves played a substantial role in the violation. This understanding, crucially, is not the same understanding that has necessarily guided online services’ compliance with FOSTA over the past five years. 

The DC Circuit court in Woodhull added that “Section 2421A(a)’s mental state requirement does not reach the intent to engage in general advocacy about prostitution or to give advice to sex workers generally to protect them from abuse.”  The court never defines what constitutes “general advocacy,” however, so platforms may still face uncertainty about what content will expose them to liability under the law. And, the fact that the court had to explicitly say that (constitutionally protected) advocacy is allowed under the law speaks to the breadth of the original language in FOSTA. 

Notwithstanding the DC Circuit’s welcome narrowing of FOSTA, in practice, the court’s reassurances about allowing advocacy around sex work are unlikely to change the chilling impacts of the law. 

FOSTA Case Law 

What impact will Woodhull have on cases relying on FOSTA for criminal prosecutions or to seek civil restitution? Likely not much, in large part given how infrequently the law has been used.  

Between its passage in 2018 and August 2021, federal prosecutors brought a single case under FOSTA: United States v. Martono (CitygXGuide), for promotion and facilitation of prostitution. (Martono pled guilty in August 2021.) By comparison, during the same time, federal prosecutors brought eight cases against online platforms using existing sex trafficking laws. The reason for this disparity is unclear. It may be because the law was still relatively new, or because prosecutors have had success using racketeering and money laundering charges against those who control online platforms in the past. 

A small number of civil suits have attempted to hold online platforms liable under the FOSTA exception to immunity under Section 230. In assessing these cases, courts have ranged in their interpretation of the mens rea needed to bring claims against online platforms under FOSTA, particularly what the language “assisting,” “facilitating” or “promoting” means. Courts have generally interpreted FOSTA’s exception to immunity under Section 230 to apply only when the online platform itself violates § 1591 of the Trafficking Victims Protection Reauthorization Act (TVPRA). For example, in a case brought against Reddit, the court held that, to fall within the FOSTA immunity exception, the “website’s own conduct [must have] violated section 1591.” It is not enough that the website’s users participated in trafficking—the website itself must “knowingly benefit from knowingly participating in child sex trafficking.” Id. at 1142. In a case against Kik Interactive, the court came to a similar conclusion, finding that under FOSTA, plaintiffs who assert claims against online platforms need to satisfy the definitions for criminal prosecution under §1591. 

The limited use of FOSTA by prosecutors and civil claimants underscores the unequal tradeoffs created by the law: its usefulness to victims appears relatively low, while its impact on expression about sex work has been severe. 

Impact of FOSTA 

Unfortunately, the narrow reading of the language in FOSTA confirmed by the DC Circuit does not undo the five years of chilling effects of the law. The vagueness of “promote,” “facilitate,” and “assist” in FOSTA caused platforms to come down hard against any content related to sex work, and sexual content more broadly, to avoid liability. After FOSTA became law, several online platforms made changes to their terms of service to restrict content around sex work. Craigslist immediately closed its “Personals” page, writing, “US Congress just passed HR 1865, ‘FOSTA’, seeking to subject websites to criminal and civil liability when third parties (users) misuse online personals unlawfully. Any tool or service can be misused. We can’t take such risk without jeopardizing all our other services, so we are regretfully taking Craigslist personals offline.” After welcoming it for years, Tumblr banned nudity in December 2018, including “photos, videos, and GIFs of genitalia and female nipples, as well as any visual depictions of sex acts” in its terms of service.  

Changes in terms of service often restrict not only activities directly involving sexually oriented materials, services, or content, but also users who have a previous connection to sex work (even if the user is not using the platform in a sex work context). Hacking//Hustling reported that sex workers who shared original tweets about Black Lives Matter from an account where they also post about sex work were significantly more likely to suffer platform policing, with 44.30% of respondents saying they noticed a difference in the visibility of their content, posts, or profile since the end of May 2020, and 51.9% saying they had lost access to financial technology (e.g. PayPal, Venmo, Square Cash, etc.). In the same survey, 82.5% of respondents who are both sex workers and “activists, organizers, or protestors” say they “avoided posting content for fear of being kicked off.”  

Online forums like Reddit and other social media sites played an important role in not just providing sex workers with a way to talk about their work and advocacy, but also as a safety measure. Sex workers could screen their clients online instead of having to make on-the-street assessments. Sex workers could advertise for and choose their own clients instead of relying on pimps to coordinate their work. Perhaps most importantly, sex workers could also talk to each other about dangerous clients on communal message boards. Now, that entire infrastructure is gone – and sex workers are paying the price with their physical safety

Some might hope that the court’s narrow interpretation of FOSTA would encourage platforms to start allowing some of this content and activity again – especially since the Supreme Court confirmed in Twitter v. Taamneh that aiding and abetting is a stringent standard to apply to the operation of an online service. But even with narrower interpretations of the law, platforms have little incentive to lighten up on their enforcement of sex-work-related content. They’ve already updated their terms of service to ban this type of content and trained their classifiers and content moderators to remove it. For many of the large platforms, sexual content makes up such a small share of the overall content (and thus profits) on the sites—just 0.06% of the content on Facebook in Q4 2022 related to nudity or sexual activity—that it’s not worth even a diminished liability risk for platforms to change their policies and practices to again allow sexual content, or sex work-related advocacy, on the platform. 

And even if they wanted to, the tools at platforms’ disposal are often not sophisticated enough to detect the nuance between people willingly engaging in sex work and those who are not. Even in an in-person conversation, it might be hard to draw the line between advocating for sex work and advertising it. But that becomes very difficult with the blunt tools of content moderation. Classifiers and other automated tools that flag prohibited content are notoriously bad at understanding the context behind a post. That context becomes even harder to understand when the large language model trying to parse it was not trained primarily in that language, or when the moderator reviewing is not a native speaker (or does speak the language at all). 

And, sex traffickers do not speak in the clear, obvious language that’s easy for tools (or even human moderators) to parse. Indeed, these are bad actors looking to thwart detection. Consider how people posted pornographic images on Facebook with “breast cancer awareness” text overlayed to avoid the company’s nudity detection tools. It’s easy to imagine similar schemes to advertise sex with people who are being trafficked. A change in the mens rea standard in FOSTA may help clarify a vague law, but it doesn’t change the fundamental challenge that content moderation tools and systems are not well equipped to draw the line between “advocating” sex work and not advertising it.

Implications for Future Legislation
Combating sex trafficking is a worthy goal. But FOSTA has fallen short of making a meaningful difference in stopping sex trafficking, while the harms to expression have been severe. Advocates of other bills that seek similar exceptions to Section 230 should take note. Vague language about the “promotion” and “facilitation” of illegal activity via online content hosting is very likely to be interpreted by courts as effectively an “aiding and abetting” standard, and draft legislation should adopt the latter standard to save everyone extended court proceedings—and to mitigate some of the chilling effect that laws like FOSTA have on online services’ willingness to host users’ speech. If legislative proposals do mean something different than aiding and abetting, they should spell out the proscribed activity and applicable standard clearly so an online platform knows what constitutes a violation of the law. Otherwise, well-meaning but vague standards will again create the cycle of over-policing speech to avoid murky liability risks.