Trump opposition to AI rules undercuts industry’s calls for a slowdown

After the leaders of Anthropic, OpenAI and other artificial intelligence companies spent the weekend warning of the dangers of advanced AI and calling for a slowdown in model development, President Donald Trump filled up his social media feed on Monday with posts undermining those efforts.

“The only control or ‘guardrails’ that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT, and the U.S.A. has that, in spades!” Trump wrote as part of an avalanche of posts slamming industry executives and critics.

Trump’s outspoken opposition to new AI regulation creates another challenge for model leaders Anthropic and OpenAI, which are trying to simultaneously grow fast enough to justify their almost trillion-dollar valuations while also responding to intensifying concerns that their technology is advancing so fast it could spin out of human control.

Anthropic CEO Dario Amodei published an essay over the weekend urging AI companies to slow their pace of development, after safety researchers Jacob Coxon and Joe Benton resigned from his lab, citing shared fears inside the industry that AI systems could potentially kill all humans.

Amodei proposed a plan for auditing and regulating AI without “sacrificing commercial advantage,” or harming U.S. national security. OpenAI CEO Sam Altman and SpaceX’s Elon Musk took the rare step of publicly agreeing with Amodei and each other.

In a fast-growing industry that lacks much by way of federal regulation, Amodei called for model evaluators to embed with frontier AI companies, and for the establishment of safety standards among democratic countries, with coordination between democratic and authoritarian governments “to the extent this is possible.”

Amodei said Anthropic is committed to giving evaluators “employee-like access” to monitor and verify the safety of the company’s models. The third-party evaluator that Amodei cited as an example was METR, a nonprofit founded by a former OpenAI and DeepMind researcher that says its mission is to “develop scientific methods to assess catastrophic risks stemming from AI systems’ autonomous capabilities and enable good decision-making about their development.”

Only a handful of other firms are in the space, including FAR.AI, AVERI and Apollo Research.

Tom Wheeler, a former chair of the Federal Communications Commission and now a visiting fellow at Brookings, said the companies can’t be responsible for effectively regulating themselves.

“Who makes the decision about standards? Who makes the decision about who’s going to do the evaluation? Who makes the decision about enforcement?” Wheeler said. While it’s important that the people at the cutting edge of AI acknowledge “great risks,” the latest proposal for industry oversight was so vague that it amounted to an “empty vessel.”

From the media and telecom industry that Wheeler helped oversee to food and beverage, autos, pharmaceuticals and financial services, virtually every major U.S. industry is subject to stricter federal oversight. AI, which is driving the U.S. economy, is the stark exception, regulated for the most part at the state level. For example, California Governor Gavin Newsom recently signed multiple AI bills into law, including Adam’s Law, which requires AI chatbot makers to protect minors from harmful interactions.

Trump, who signed an executive order in June asking AI companies to voluntarily submit their models to the government before release, is adamantly opposing any federal regulation and insists that fears about AI amount to a “hoax.”

“Concerning AI, when, in the History of Business, did anyone see the Leaders of an Industry call for Regulation that, if strongly implemented, will drive them into oblivion and bankruptcy?” he wrote in a Truth Social post on Monday.

Meanwhile, Trump’s former AI czar, David Sacks, continued to criticize what he’s repeatedly called Anthropic’s “regulatory capture” campaign. In a weekend post responding to Amodei’s essay, Sacks, a venture capitalist with investments in AI startups, said the two leading model developers should police themselves rather than asking the government to do it.

“If the unreleased models are scary enough that you think you should slow down, I support your decision to be responsible,” he wrote.

Representatives from Anthropic and OpenAI didn’t immediately respond to requests for comment.

The Trump administration has taken the stance that anything that impedes AI development in the U.S. is to the benefit of chief adversary China, the world’s second-biggest economy. Across the private sector, the president has shown a willingness to try and hurt companies that take a stand against him, said Jessica Ji, a senior research analyst at Georgetown University’s Center for Security and Emerging Technology.

Anthropic has already experienced such treatment, getting blacklisted by the Department of Defense earlier this year, and then in June being ordered by the government to disable access to two models to comply with an export control directive.

“There are serious concerns like punishment by the Trump administration,” Ji said. “The administration is trying to incentivize them to not pause and to keep developing their models and releasing them.”

In addition to the comments from Amodei, Altman and Musk, Microsoft CEO Satya Nadella wrote on Sunday in support of “deliberate pacing,” while Demis Hassabis, the chair of Google DeepMind, wrote on X that “Dario’s essay points towards the right path forward.” 

Meta CEO Mark Zuckerberg, who’s been aggressively pushing his company’s latest models while also trying to improve his relationship with Trump, has remained mum on the matter so far.

Meta didn’t respond to a request for comment.

Third-party evaluations can work, Johri said, when it’s a project like OpenAI’s Daybreak or Anthropic’s Glasswing, which allowed some companies to access, test and provide feedback on the technologies. Partners “can help fortify the defenses before capabilities are made broadly available,” Johri said, noting that Checkmarx is part of the Glasswing program.

Those efforts only involve cybersecurity and software, Johri said, and any potential AI-safety related initiative would need a more expansive range of subject-matter experts. The potentially catastrophic risks of things going wrong only heightens the sense of urgency.

“Unfortunately, things happen fast when there’s a crisis,” he said. “I don’t know if there’s a crisis right now, but it certainly is a good discussion to be had.”

Amba Kak, co-executive director of the AI Now Institute, said in an email that the AI industry is in dire need of “enforceable government regulation” aimed at preventing harm and ensuring safety standards of the sectors where the technology is being sold.

“Failing that requirement should mean losing market access and facing the most punitive forms of liability, including for executives personally,” said Kak, a former senior AI adviser for the Federal Trade Commission. “Nothing short of that will get us the leverage we’re fundamentally missing against this industry.”

Third-party evaluators are not a sufficient replacement for independent government oversight, she said.

“We’re up against decades of impunity and opacity,” Kak said. “We need a structure that doesn’t rely on this industry’s permission to do its job.”

Timnit Gebru, founder of the Distributed AI Research Institute (DAIR), said plenty of laws already apply to AI today, and companies’ claims of existential risks are overblown.

“There is no AI exemption in existing statutes,” she said. “Fraud is still fraud. Discrimination is still discrimination.”

WATCH: Genie is out of the bottle with AI, says Crowdstrike CEO George Kurtz.