The TRUMP AMERICA AI Act Is Every Bit As Bad As You Would Expect. Maybe Worse.

Sen. Marsha Blackburn’s latest is an anti-tech omnibus, combining years' worth of dangerous policy ideas into one big, bad bill.

Sometimes you can tell a bill will be really bad just from its title. So it goes with The Republic Unifying Meritocratic Performance Advancing Machine Intelligence by Eliminating Regulatory Interstate Chaos Across American Industry Act, from Sen. Marsha Blackburn (R–Tenn.). And, boy, does it deliver on that disaster of a name, managing to combine nearly every bad tech policy idea of the past half-decade—including gutting Section 230 and creating new requirements around the suppression of sexuality online—into one massive piece of Trump-branded legislation.

The bill's title alone is asinine, even if we put the North Korea-ness meets word-salad nature of it aside. Following the normal rules of making acronyms, it would be the TRUMP AMIERICA (or perhaps AMIBERICA) AI act, though Blackburn is throwing rules to the wind and referring to it as the TRUMP AMERICA AI act.

If only the problems stopped there!

Alas, Blackburn is serving up a cornucopia of proposals that could throttle free speech and free markets online. An anti-tech omnibus, if you will, sold as a simple AI regulatory scheme.

Techdirt's Mike Masnick calls it a "massively destructive internet policy overhaul masquerading as AI legislation." It "would change nearly every US government policy regarding how the internet works, tackling AI, Section 230, copyright, and a bunch of other nonsense all in one bill."

'Duties of Care' All Around 

Masnick has a nice rundown of the bill's myriad flaws, which include instituting a "duty of care" for AI developers to "prevent and mitigate foreseeable harm to users" (per Blackburn's summary of the bill). This duty would be enforced by the Federal Trade Commission (FTC).

"This is one of those things that I'm sure sounds good to folks, but as we've explained over and over again this kind of 'duty of care' is basically an anti-230 that would do real damage," writes Masnick.

It's basically just an invitation for lawyers to sue any time anything bad happens and someone involved in the bad thing that happened somehow used an AI tool at some point.

And then you have to go through a big expensive legal process to explain "no, this thing was not because of AI" or whatever. It's just a massive invitation to sue everyone, meaning that in the end you have just a few giant companies providing AI because they'll be the only ones who can afford the lawsuits.

And just in case that didn't allow for enough ways to attack AI companies, another section of the bill would enable "the U.S. Attorney General, state attorneys general, and private actors to file suit to hold AI system developers liable for harms caused by the AI system for defective design, failure to warn, express warranty, and unreasonably dangerous or defective product claims."

Blackburn—who was once a proponent of light-touch regulation when it came to the internet—has also worked elements of the Kids Online Safety Act (KOSA) into the TRUMP AMERICA AI Act.

It will require certain social media platforms, video games, stream services, and messaging applications "to implement tools and safeguards to protect users and visitors under the age of 17 to protect children from sex trafficking, suicide, and other abuses," per Blackburn's summary. As with KOSA, this requirement is promoted in a way that sounds unobjectionable—admirable, even—but would, in effect, require companies to suppress massive amounts of content, weaken privacy protections, and more.

"This section generally requires covered platforms to exercise reasonable care in the design and use of features that increase minors' online activity to prevent and mitigate harm to minors (e.g., mental health disorders and severe harassment)," the summary says.

Enterprising lawyers can easily argue that all sorts of things contribute to mental health issues in their young clients, enabling lawsuits over generally unobjectionable (or, at the very least, totally legal) speech and neutral platform features. The biggest tech companies may be able to fight these, but all but the behemoths would be forced to preemptively ban a bunch of speech in order to avoid potential lawsuits.

Ushering Pro-Conservative Bias Into AI

Section 11 of Blackburn's bill is promoted as combating "the consistent pattern of bias against conservative figures demonstrated by Big Tech and AI systems." But, in practice, it could require AI systems to have a pro-conservative slant—at least as long as President Donald Trump or other Republicans are in power.

The bill would set up "audits of high-risk AI systems to undergo regular bias evaluations to prevent discrimination based on protected characteristics, including political affiliation."

Presumably, federal agencies would be tasked with conducting these audits, which could leave it up to political appointees—not exactly a notoriously unbiased bunch—to judge what does and doesn't count as bias against a particular political group. How long before AI developers have to tailor their systems to spitting out politically favorable results?

The effect of this section could be somewhat blunted by the fact that it only applies to "high-risk" systems, which Blackburn's summary describes as "those that could pose significant risks to health, safety, rights, or economic security, including those in education, employment, law enforcement, or critical infrastructure." But without a more precise definition, it's hard to say how this would shake out or what it would mean for the sorts of general AI systems used by consumers.

Making Big Tech Less User-Friendly

During the heyday of federal antitrust hearings about Big Tech, the idea of ending "self-preferencing" got a lot of play. Self-preferencing refers to tech companies using their services to promote or favor their other services, and for some reason, lawmakers are convinced that it's a scourge.

But self-preferencing comes with a lot of perks for tech users, not just for the companies involved. It means that when you Google a particular place or business, Google will automatically place a map of this location near the top of the search results. It means that Amazon will perhaps show you more products eligible for free shipping with a Prime membership—something Prime members want!—than products where shipping costs extra. And so on.

The TRUMP AMERICA AI act would stop "systemically important platforms"—defined as including, but perhaps not limited to, "platforms with subscribers or monthly active users in the United States not less than 34% of the population of the United States"—from engaging in "self-preferencing or steering users to products or services offered by the platform operator," per Blackburn's summary.

In effect, it would make Big Tech less user-friendly in the name of protecting us from Big Tech.

A Backdoor to Banning Porn on Big Tech Platforms  

A line tucked near the bottom of Blackburn's summary says that the bill would prevent "systemically important platforms from disseminating sexual material harmful to minors."

It's cloaked in euphemistic language: "sexual material harmful to minors" sure sounds like something very bad, like it might be referring to child pornography or other forms of illegal imagery.

But we've seen, in myriad state laws targeting material harmful to minors, that this term can be used very broadly, encompassing not just any and all pornographic photos and videos but also written erotica, literature that describes sexual relationships, stories centered on gay and transgender characters, and so on.

A requirement that big tech platforms ban "sexual material harmful to minors" would almost certainly mean that they must filter out anything that could be considered porn and perhaps much more.

Gutting Section 230 

One of the most worrying bits of the bill concerns Section 230 of the Communications Decency Act. Blackburn's bill would "establish a 'Bad Samaritan' carve-out that would deny immunity from civil liability to platforms that purposefully facilitate or solicit third-party content that violates federal criminal law."

Of course, Section 230 is already inapplicable to violations of federal criminal law. A company can't break federal law and claim that Section 230 lets them do it.

So what's the true aim here? I think Masnick frames the issue pretty well:

Right now, 230 lets platforms get frivolous lawsuits dismissed quickly at the motion to dismiss stage. This change would force every platform to go through lengthy, expensive litigation to prove they weren't "facilitating" (an incredibly vague term) or "soliciting" third-party content that violates federal criminal law.

That's gutting the main reason Section 230 exists. Instead of quick dismissals, you get discovery, depositions, and trials, all while someone argues that because your algorithm showed someone a post, you were "facilitating" whatever criminal content they claim to find.

Slippery words like "facilitate" and "solicit" give authorities a lot of leeway to punish tech companies for activities we generally think of as non-criminal, free-market, or speech-facilitating activities.

Chatbot Regulation, State Law Preemption, and Other Changes 

The bill would put into policy Trump's desire to ban states from passing their own AI regulation. Earlier this month, the president issued an executive order seeking to stop states from passing certain sorts of AI regulation so the country could have, instead, a "national framework"—though the order can't actually create said framework or outright ban states from passing their own laws. Congress can, however. And Blackburn's bill would preempt state AI laws in several arenas.

Blackburn's summary also lists a huge array of other changes the TRUMP AMERICA AI Act would enact. Some of these summaries are relatively vague—for instance, Section 8 is merely described as "establish[ing] requirements for companies providing AI chatbot and companion services to protect kids."

One section would require "interoperability for systemically important platforms, which include platforms with subscribers or monthly active users in the United States not less than 34% of the population of the United States." Interoperability is one of those ideas that may sound nice in theory but presents huge technical challenges and security risks.

Several sections seem designed to upend copyright laws, by ignoring concepts like fair use, satire, and parody. There's a bit that would create "a federal right for individuals to sue companies for using their data (personal, copyrighted) for AI training without explicit consent" and another that would "hold individuals or companies liable if they produce an unauthorized digital replica of an individual in a performance." Yet another section would deem "derivative works generated, synthesized, or produced by an AI system without authorization as infringing works, which would be ineligible for copyright protection."

Does It Have Legs? 

The bill hasn't even been formally introduced yet, let alone attracted official cosponsors, so it's hard to say how Blackburn's colleagues will treat the bill. But it seems clear that the measure's title has been calculated to attract Trump's endorsement, which could translate to a lot of Republican lawmakers falling in line, too.

Blackburn's announcement of the TRUMP AMERICA AI Act is also steeped in MAGA flattery and rhetoric. The bill would "codify President Trump's executive order to create one rulebook for artificial intelligence," it says.

"I look forward to introducing the TRUMP AMERICA AI Act in the new year to create one federal rulebook for AI to protect children, creators, conservatives, and communities across the country and ensure America triumphs over foreign adversaries in the global race for AI dominance," said Blackburn.

---

More Sex & Tech News

New post!

There was a lot of innovation in medicine and biomedical research this year, and I've tried to summarize the biggest ones in this blogpost.

Medical breakthroughs in 2025:https://t.co/A18geeO7bt pic.twitter.com/7HbErpsfwY

— Saloni (@salonium) December 28, 2025

Patient "states he has a foreign body in his rectum that is vibrating. He states he was with a girl last night and doesn't remember much." Using data from the U.S. Consumer Product Safety Commission's emergency room visits database, Defector has compiled a list of things people got stuck in their rectums and genitals in 2025.

New York passes an immunity bill. The bill "provides immunity from prosecution for certain individuals engaged in prostitution who are victims of or witnesses to a crime and who report such crime or assist in the investigation or prosecution," per the legislative summary. "This law recognizes that safety must be prioritized over punishment," said Decriminalize Sex Work Legal Director Melissa Broudo. "It is a vital and common sense public safety measure that strengthens law enforcement's ability to identify, investigate, and convict perpetrators of violence and trafficking."

Did China just ban sexting? "The Chinese government has banned the sharing of 'obscene' content in private online messages and increased the penalties for spreading pornographic material," reports The Washington Post. "While the revision will target the dissemination of pornography and exploitative images," the new regulation "may also mean that consensual sexting could also be dragged into China's legal system."

Lol: The URLs trumpkennedycenter.org and trumpkennedycenter.com are owned by comedy writer Toby Morton, who predicted the renaming of the D.C. performing arts institution (it will become the "The Donald J. Trump and The John F. Kennedy Memorial Center for the Performing Arts") and snapped up the web domains in advance.

---

Today's Image

Washington, D.C. | 2017 (ENB/Reason)