The Senate Commerce Committee this week advanced several bills that would radically restrict free expression, innovation, and privacy online. Most policymakers sincerely believe these bills will keep kids safe online, but the bills are dangerous attacks on online speech. The committee took up measures on both online platforms and AI; this article focuses on the two bills aimed at social media, gaming, and communications platforms: the SCREEN Act and the Kids Online Safety Act (KOSA).
The SCREEN Act
The SCREEN Act won a 15–13 committee vote but was not reported out because too few senators were physically present to meet the quorum requirement, so we will likely see it again. It aims to keep minors from accessing pornography. But any platform hosting any potentially pornographic content must verify the age of all its users. The bill also targets virtual private network (VPN) users.
While short in size, the list of problems with this bill is long. First, there is the recurring problem with age verification regimes. Age verification isn’t just something minors must do; every user must do it. The result is that platforms will need to collect massive amounts of identifying information about nearly all of their users. This mass of data creates an incentive for hackers and makes the amount of information released in data breaches even larger.
And this isn’t just an abstract fear, as there have been major breaches of companies that verify age or identity. For example, in 2025, Discord’s age-verification firm was hacked, releasing 70,000 government IDs, along with other identifying information and emails. The women-only dating safety app Tea was breached the same year, exposing 72,000 identifying images and 1.1 million private messages.
These breaches create tangible harms. Exposed personal data fuels identity theft that damages victims’ finances and well-being—especially for kids, who may not discover it until adulthood, by then saddled with damage to their future opportunities.
But this increased risk of data breaches also threatens free speech. Minors shut out of a platform lose access to information, and the ability to express themselves—but adult speech is also threatened. Think about all the sensitive topics that users may engage with online: sexual, health, religious, or political speech, groups for former addicts or those in abusive situations, government whistleblowers or critics, and plenty more. When engaging with these topics, many users want anonymity, as users of Tea did. But age verification threatens that anonymity, as every speaker must be identified to access the platform. So whenever anyone speaks about or accesses sensitive information, they will do so knowing that a data breach or hack could identify them and their speech. This chills the speech of adults.
But the SCREEN Act is just about porn, one might object. Yes, that is its stated objective. But the way it is drafted defines nearly every single platform on the internet as a platform carrying pornography. If a platform carries a single piece of pornographic content, it may be required to age-verify all its users accessing all sorts of non-pornographic and First Amendment-protected speech.
While Senator Cruz argued in the markup that this is similar to what has passed in the House and the Texas bill that was deemed constitutional by the Supreme Court in FSC v. Paxton, those bills applied to platforms where more than one-third of the content was sexual material harmful to minors. While there were plenty of other problems with those bills, the SCREEN Act sets such a low threshold that it covers nearly all online platforms. Even though I disagree with the ruling in FSC v. Paxton, it’s hard to imagine that it justifies any single instance of potentially pornographic content triggering a vast age-verification requirement for a platform.
And the SCREEN Act doesn’t stop there by also attacking virtual private network (VPN) usage. As I’ve written about previously, VPNs are tools that many individuals and businesses use to enhance their online security and privacy. But VPNs also route your IP address through another location, making it appear that you are somewhere that you are not. This is an effective way to circumvent age-verification requirements. Indeed, each time a US state or a foreign country tries to implement some new online speech regime, we quickly see VPN usage skyrocket.
Rather than admit the underlying policy is flawed, policymakers double down: the SCREEN Act requires platforms to age-verify anyone arriving via a known VPN IP address, further undermining privacy tools. This provision also reaches beyond US borders. If someone in another country accesses a platform through a VPN, the platform must age-verify that foreign user before granting access, simply because the traffic arrives via a known VPN IP address. Americans rightly criticize foreign governments for imposing their rules on Americans, but the SCREEN Act would impose American rules on those outside of America.
Kids Online Safety Act (KOSA)
The Senate version of KOSA passed without objection from the Commerce Committee and is similar to the flawed House version. That said, KOSA sponsor Senator Blackburn proudly declared the Senate version to be far stricter, calling the House version “toothless.” Senator Blackburn is right that the Senate bill is a more aggressive assault on free expression and innovation.
The Senate KOSA applies to nearly all online platforms—video-streaming services, video games, and messaging tools other than email, video conferencing, or texting. While both versions of KOSA create various requirements for platforms to follow with regard to content and design features, the Senate KOSA is built around creating a “duty of care.” Platforms are required to prevent and mitigate a series of “reasonably foreseeable” harms that a “reasonable and prudent person” would agree that the design features of a platform contribute to, including:
(1) Eating disorders, substance use disorders, and suicidal behaviors.
(2) Depressive disorders and anxiety disorders when such conditions have objectively verifiable and clinically diagnosable symptoms and are related to compulsive usage.
(3) Patterns of use that indicate compulsive usage.
(4) Physical violence or online harassment activity that is so severe, pervasive, or objectively offensive that it impacts a major life activity of a minor.
(5) Sexual exploitation and abuse of minors.
(6) Distribution, sale, or use of narcotic drugs, tobacco products, cannabis products, gambling, or alcohol.
(7) Financial harms caused by unfair or deceptive acts or practices (as defined in section 5(a)(4) of the Federal Trade Commission Act (15 U.S.C. 45(a)(4))).
The FTC is empowered to enforce the law, with each violation being considered an unfair or deceptive act or practice. Each state attorney general may also bring civil actions to enforce portions of the law if the FTC does not act.
While there are significant differences from the House version, the biggest similarity is that both pose a major threat to free expression. To start with, Senate KOSA will likely require mass age verification across all platforms if companies are to fulfill their duty of care to minors. The harms of age verification, discussed above, apply here as well.
But KOSA goes much further by directly attacking speech. Looking again at the harms that KOSA targets, many of them are categories of speech: eating disorders, suicidal tendencies, violent or harassing content, drug, tobacco, or alcohol content, etc. Yet KOSA does nothing to punish those who post illegal content; it merely makes platforms responsible for the speech of others.
Beyond illegal content, the bill also covers harms like harassment or content addressing issues like suicide that, while potentially unsettling, are often protected speech. For example, many on the left would consider misgendering or deadnaming transgender individuals to be harassment. Twitter held such a view and banned such speech—a stance that became one of the grievances Elon Musk cited before buying the platform. Many on the right would find various forms of anti-Christian or anti-Jewish speech to be a form of harassment. Requiring platforms to tackle a broad category of “online harassment” is simply a demand that platforms censor protected speech.
Ultimately, KOSA is all about regulating speech while being framed as focused on “the design of platform.” When members of Congress and other promoters of KOSA speak, they frequently talk about addiction and compulsive use that is causing various harms, a causation that is far from settled science. But then the jump is quickly made to harmful content. Senator Capito stated in the hearing that “we can better safeguard our youth and filter out violent, sexual, and deceptive content that they may encounter.”
Imagine a world where kids mostly used social media to study better, eat healthier, or pick up hobbies like gardening or classical music. If they were endlessly scrolling that content, pinging each other about it, and getting recommendations for more, no one would be outraged, and no one would be drafting bills to hold tech companies accountable. The issue isn’t design features or compulsive use—it’s really about the content kids consume.
The reason why policymakers drafted these bills to focus on design features, though, is because blatant regulation of the content of speech is highly constitutionally dubious. Content-based regulation faces the greatest scrutiny in the courts, so bills like KOSA frame the problem as a product-design issue to circumvent such constitutional infirmities. But it still all comes down to speech.
This policy is bad enough, but KOSA gets even worse by enforcing a duty of care on platforms. The duty of care is a legal responsibility for platforms to design their products to reasonably prevent or mitigate the harmful categories listed above. Thus, platforms must remove speech or features that could contribute to or could be related to those harms. But this incredibly broad liability creates a moderator’s dilemma: moderate too little and the company is liable. Moderate too much, and you silence lots of important speech, but are not liable.
For example, KOSA makes platforms responsible for preventing their products from contributing to children’s depression or anxiety—but what kind of content should a platform allow? Political, religious, or climate content; pro-LGBTQ and anti-LGBTQ content; posts showing others doing better than you—or worse; violent video games and even placid farming simulators—all of it could be said to contribute to depression or anxiety.
Again this isn’t just hypothetical. When the UK implemented its similar “Online Safety Act” discussions of the conflicts in Ukraine or Gaza were blocked. Content as varied as music streaming services and politicians advocating for their political party were caught up in efforts to keep kids safe online.
And the duty of care can change depending on the administration, FTC commissioners, or state attorneys general. What is responsible and reasonable one day and in one jurisdiction will be irresponsible and unreasonable on another day or in another jurisdiction. No company can effectively operate with such a sword of Damocles hanging over its head. Facing such a dilemma, most platforms will restrict vast amounts of content and platform features. For many smaller platforms and startups, this will simply end their business. While targeted at kids, this will vastly restrict the tools and speech that are available to many Americans.
Empowering Americans, Not Government
Many Americans and policymakers support these proposals because they honestly believe they will keep kids safe. These bills demand that social media, gaming, and communications platforms police the internet for kids—acting as parents, serving only content they deem appropriate, with the government stepping in when it decides they’ve failed.
But platforms and governments make bad parents. They simply can’t do what is best for each child because many parents disagree on what is appropriate for their children. Some embrace technology, others shun it. Some are progressive, some are libertarian, some are conservative, some are apolitical—and they all have different ideas about what is good for kids.
Rather than censoring and chilling legitimate speech, a far more effective and tailored approach is to empower Americans: teaching children how to be safe online and giving parents the resources to navigate an ever-evolving set of online technologies.
The SCREEN Act and KOSA fail on both counts: they protect neither kids nor free speech.
