The Trump administration on Friday unveiled its new legislative blueprint for AI regulation, and the seven-point plan features a clear message: The federal authorities ought to keep away from many AI rules past a set of kid security guidelines, and it ought to bar states from messing with the “nationwide technique to realize international AI dominance.”
The plan advises Congress to guard minors utilizing AI providers with extra safeguards and take motion to try to stop electrical energy prices from spiking because of AI infrastructure. It encourages “youth improvement and abilities coaching” to spice up familiarity with AI instruments, with out a lot additional element. But it surely suggests taking a wait-and-see method as to whether coaching AI fashions on copyrighted materials with out permission is authorized, and it maintains a long-running Republican push to restrict whether or not states can enact their very own AI legal guidelines.
The complete doc and all its provisions, nevertheless, will solely take impact if Congress adopts them into laws and passes them into legislation.
The Trump administration blueprint encourages passing legal guidelines much like the Take It Down Act — which was signed into legislation in Might 2025 and bars nonconsensual AI-generated “intimate visible depictions,” requiring sure platforms to quickly take away them. The doc is also pro-age verification, suggesting that Congress “set up commercially cheap, privateness protecting, age assurance necessities (resembling parental attestation) for AI platforms and providers prone to be accessed by minors.” Age-gating is controversial from a privateness standpoint and has a variety of potential surveillance implications. It proposes different youngster safety measures like limiting the power for AI fashions to coach on minors’ information and limits to focused promoting primarily based on their information. (The doc doesn’t search to ban these practices for youngsters’s information, simply restrict them.) On the identical time, it states that Congress “ought to keep away from setting ambiguous requirements about permissible content material, or open-ended legal responsibility, that might give rise to extreme litigation.”
Within the age of deepfakes, when AI-generated movies are trying extra actual than ever and a pretend video of a politician can immediately propagate international conspiracy theories, the brand new coverage blueprint seeks to “take into account establishing a federal framework defending people from the unauthorized distribution or business use of AI-generated digital replicas of their voice, likeness, or different identifiable attributes.” (That would imply lastly making a federal likeness legislation.) But it surely additionally says lawmakers ought to present “clear exceptions” for parody, information reporting, satire, and different First Modification-protected use circumstances.
The blueprint additionally discourages Congress from taking over AI copyright points. “Though the Administration believes that coaching of AI fashions on copyrighted materials doesn’t violate copyright legal guidelines, it acknowledges arguments on the contrary exist and subsequently helps permitting the Courts to resolve this difficulty,” it says. “Congress mustn’t take any actions that might influence the judiciary’s decision of whether or not coaching on copyrighted materials constitutes honest use.”
In one other part, the blueprint raises issues about large-scale scams and fraud which are more and more powered by AI, stating that Congress ought to “increase current legislation enforcement efforts to fight AI-enabled impersonation scams and fraud that focus on susceptible populations resembling seniors,” though no additional particulars are offered.
The Trump administration continued leaning into the pro-federal, anti-state method to AI regulation that it’s been selling (thus far unsuccessfully) for almost a yr. The blueprint says Congress ought to “preempt state AI legal guidelines that impose undue burdens” and keep away from “fifty discordant” requirements for firms, including that states “shouldn’t be permitted to manage AI improvement, as a result of it’s an inherently interstate phenomenon with key overseas coverage and nationwide safety implications.” Different authorized protections for AI firms have been baked in, too, resembling the concept states shouldn’t be allowed to “penalize AI builders for a 3rd social gathering’s illegal conduct involving their fashions.” However within the child-privacy part, the doc does enable states some restricted wiggle room, stating that Congress shouldn’t preempt states from “imposing their very own usually relevant legal guidelines defending youngsters, resembling prohibitions on youngster sexual abuse materials, even the place such materials is generated by AI.” The allowance comes after quite a few figures from each events expressed concern about overturning native youngster security legal guidelines, together with almost 40 attorneys basic for US states and territories.
The general purpose, as in earlier Trump administration proposals, is dashing AI improvement. “The US should lead the world in AI by eradicating obstacles to innovation [and] accelerating deployment of AI purposes throughout sectors,” the doc states, including that Congress ought to discover methods to make federal datasets out there to AI firms and teachers in “AI-ready codecs to be used in coaching AI fashions and programs.” It didn’t specify which forms of federal datasets it sought to make publicly out there for AI coaching. The plan additionally definitively solutions a long-asked query in AI regulation — whether or not there must be one federal physique liable for AI regulation or whether or not AI regulation must be left to every sector — and says that Congress “mustn’t create any new federal rulemaking physique to manage AI”; as a substitute, it says, it can “assist improvement and deployment of sector-specific AI purposes via current regulatory our bodies with material experience.”
President Trump signed an government order final July in search of to stop “woke AI” by banning authorities companies from utilizing fashions that “integrated” subjects like systemic racism. He just lately ordered all companies to blacklist the “Radical Left AI firm” Anthropic for setting limits on navy use of its fashions, one thing Anthropic alleges violates its First Modification rights. On the identical time, the blueprint states that the federal government “should defend free speech and First Modification protections, whereas stopping AI programs from getting used to silence or censor lawful political expression or dissent.” It goes additional to say that Congress ought to explicitly forestall the federal government from “coercing” AI suppliers “to ban, compel, or alter content material primarily based on partisan or ideological agendas” — and that within the occasion that authorities companies censor expression on AI platforms or dictate the data they supply, then Congress ought to present a means for People to “search redress.”
Final month, we noticed the primary bipartisan effort to handle larger utility payments in communities with information facilities close by, and the brand new AI coverage framework appears to handle these issues on each side of the aisle, saying that Congress ought to discover methods to ensure that “residential ratepayers don’t expertise elevated electrical energy prices because of new AI information middle building and operation.” However, it says, Congress ought to streamline federal permits for information middle building and operation, making it simpler for AI firms to and make it simpler for “develop or procure on-site and behind-the-meter energy technology” — that means that information middle building ought to nonetheless be full-speed-ahead, however neighborhood members shouldn’t need to actually pay the value on their month-to-month payments.
Comply with subjects and authors from this story to see extra like this in your customized homepage feed and to obtain e mail updates.
- AIShut
AI
Posts from this matter might be added to your every day e mail digest and your homepage feed.
Comply withComply with
See All AI
- EvaluationShut
Evaluation
Posts from this matter might be added to your every day e mail digest and your homepage feed.
Comply withComply with
See All Evaluation
- CoverageShut
Coverage
Posts from this matter might be added to your every day e mail digest and your homepage feed.
Comply withComply with
See All Coverage
- ReportShut
Report
Posts from this matter might be added to your every day e mail digest and your homepage feed.
Comply withComply with
See All Report
