American Compass Points the Wrong Way on Online Regulation

C hris Griswold, the policy director at American Compass, recently argued that the Kids Online Safety Act (KOSA) does not threaten free speech as many conservatives and libertarians have warned. Griswold quoted my argument that Americans should consider the online world to be an extension of “real life” subject to established constitutional protections rather than an accessory to life open to extra-constitutional restrictions. He responded: “American social and legal norms not only permit but require us to keep kids out of dangerous environments like nightclubs, protect environments like playgrounds that are meant to be safe for kids, and prevent kids’ access to harmful products like alcohol.”

He chooses the wrong analogy. Going online resembles walking out the front door more than entering a nightclub. The internet — much of which flows through social-media sites these days — is not a place one frequents only for a night out. The digital world hosts an ever-increasing array of nearly all aspects of life — work, education, entertainment, socializing, communication, and more.

Speech-limiting proposals like KOSA invite First Amendment scrutiny while forbidding minors to buy alcohol does not. Further, as Griswold’s op-ed elides, KOSA and similarly conceived bills threaten the speech rights and privacy of all internet users, not just minors.

Besides this larger mistake, Griswold’s two primary arguments about KOSA — that its duty of care is clearly defined and narrowly tailored and that there is little chance bureaucrats will abuse its provisions — wilt under scrutiny.

KOSA would impose on social-media platforms a duty of care to take “reasonable” steps to protect minors from online harm. To dodge the liability it would impose, platforms would need to remove or suppress their users’ online speech when it could contribute to any of the listed harms. Affected speech inevitably would include — if not primarily consist of — individual Americans’ posts, not just linked news stories, as Griswold suggested.

Griswold said KOSA is “narrowly tailored” and “carefully target[ed at] only the most serious and well-documented health risks of social media.” However, the proposal enumerates roughly 15 types of harm, ranging from the uncomfortable and subjective (e.g., anxiety-inducing) to objective and abject horrors (e.g., sexual abuse and suicide).

What’s more, how platforms should carry out their new duty remains entirely unclear. KOSA does not outline how to balance free expression and children’s safety or how censorship-happy platforms must be to fulfill their obligations under the law. The “reasonableness” standard the bill employs is famously malleable. Many types of content could contribute to a child’s anxiety or depression; what might inflame one teen’s insecurities might not impact his fellows. For example, banal fitness-related content could inspire one young person to exercise more while aggravating another’s preexisting eating disorder.

The bill leaves these questions to the discretion of platforms and, ultimately, the bureaucrats who would enforce it. Thus, the bill would, if enacted, incentivize litigation-averse platforms to preemptively remove a large amount of speech to avoid even the faintest whiff of noncompliance.

Griswold’s pooh-poohing of KOSA’s potential to suppress speech slams headlong into myriad inconvenient facts. He neglected to report that the current FTC (tasked to enforce the bill) odiously targeted journalists who published the Twitter Files, a reporting series that exposed the Biden administration’s efforts to suppress disfavored speech online. He also omitted the fact that the bill allows for enforcement by state attorneys general as well as the federal watchdog, which increases the chances of censorship cutting across the political spectrum.

Griswold writes as if everybody would agree on KOSA’s implementation. He has little basis for this optimism. For example, the Left and Right diverge wildly on the prudence of supplying hormone treatment to transgender-identifying minors. The left says withholding such drugs causes “depression” and “suicidal behaviors” (two enumerated harms in KOSA’s duty of care). Alternatively, the Heritage Foundation has endorsed employing KOSA to “[keep] trans content away from children.”

Consider the Left’s myriad, unflaggingly creative arguments that speech constitutes or materially contributes to violence. Extensive reporting and legal discovery revealed the Biden administration’s hard pressure on social-media platforms to remove speech that deviated from the public-health establishment’s narratives regarding Covid-19. The administration justified this malfeasance as necessary to preserve public health. President Biden charged platforms that failed fully to accede to his demands with “killing people.” These efforts did extreme damage to American free speech. No conservative should favor further emboldening left-wing bureaucrats and the Disinformation Brigade.

For years, the Left has asserted that “speech is violence” and that such ordinary phrases as “America is the greatest country in the world” may constitute unconscionable psychological warfare against minority groups. Democratic officials (or social-media executives) could easily presume that non-leftist speech on these issues falls under KOSA’s enumerated harms of “anxiety,” “online bullying,” or “harassment.” The bill would give platforms’ anti-conservative biases Washington’s imprimatur.

Griswold asserts that, under the FTC’s controversial 2022 Section 5 policy statement, the agency already claims the authority to intervene in online content moderation. But here, he has confused two distinct FTC authorities. The policy statement articulates the agency’s view of its power to police Unfair Methods of Competition (UMC, which concerns antitrust violations), not its power to police Unfair and Deceptive Acts and Practices (UDAP, which concerns consumer protection). KOSA would grant new UDAP authority to the FTC. The FTC’s views in the UMC policy statement have no bearing on cases that involve children’s online safety (i.e., UDAP cases). And even if the Section 5 statement were relevant, its cavernous interpretation of FTC authority would clash with decades of settled precedent and would likely be struck down if scrutinized in court.

Griswold’s positive reference to the Federal Communications Commission’s (FCC) power to regulate broadcast media harms his own case, because the FCC weaponized this power over decades to suppress heterodox and politically inconvenient speech.

Finally, Griswold stated correctly that the judiciary generally scraps laws that violate speech rights. Accordingly, courts have spent decades invalidating laws that (like KOSA) pursue children’s online safety in unconstitutional ways. These cases include Reno v. ACLU (1997), multiple cases eviscerating the Child Online Protection Act of 1998, and Brown v. Entertainment Merchants Association (2011), as well as opinions this year enjoining laws in California, Arkansas, and Texas.

KOSA will meet a similar end in the courts — unless Congress makes use of Chris Griswold’s foresight and declines to enact it in the first place.