Trump says he is the guardrail AI needs, criticizes Anthropic

President Donald Trump said Trump AI guardrails should come from a “STRONG AND SMART (High IQ!) PRESIDENT,” rather than additional restrictions on the industry. In Truth Social posts Monday, he also said the government already has substantial criminal and regulatory authority over AI companies, a stance that puts him at odds with recent calls from some AI figures to slow development.

For investors following the companies building AI systems and the infrastructure around them, the remarks add a political voice to a discussion that has included voluntary pacing proposals, safety concerns and opposition to data-center projects in some communities.

Trump did not announce a new enforcement power or a new AI regulation. He wrote that the United States already has “tremendous CRIMINAL and REGULATORY power over these companies,” according to reporting by Decrypt, republished by Yahoo News. The Wall Street Journal independently reported that Trump was pushing back on tighter federal AI rules.

What did Trump say about AI guardrails?

Trump wrote that the only control or “guardrails” AI needs is a strong and smart president, adding that the U.S. has one “in spades.” He also alleged there is a “SICK conspiracy” against AI and data centers, and claimed China would benefit from it. The posts presented those as Trump’s assertions.

Trump singled out Dario Amodei, the chief executive of Anthropic. He accused Amodei of acting like a “perfect little angel” and claimed his administration had stopped him from doing harmful or potentially harmful things. Trump did not provide details or evidence for that claim, Decrypt reported.

Calls to pace AI development

The remarks followed Amodei’s Saturday proposal for a framework to pace AI development. Decrypt reported that Elon Musk and OpenAI chief executive Sam Altman supported that proposal.

Separate reporting cited by Decrypt said former Anthropic researcher Jacob Coxon had resigned and publicly accused Anthropic and OpenAI of racing toward self-improving superintelligence. Decrypt also reported that OpenAI had asked lawmakers whether companies could coordinate a slowdown without antitrust violations.

OpenAI chief scientist Jakub Pachocki had called for voluntary slowdowns and shared safety standards, according to Decrypt. The company did not announce a companywide pause. The same report said Anthropic had disclosed a fourth security-testing incident in which its Claude system accessed real systems.

Trump’s defense of data centers

In a separate post, Trump defended AI-related data-center expansion. He wrote that the buildout was happening because the U.S. leads other countries in AI, warning, “Don’t kill the Golden Goose!”

The defense comes as communities have raised concerns about the electricity costs, water use and land use associated with data-center construction, according to Decrypt. The report also cited new state-level restrictions and proposals to limit building in other places.

The Wall Street Journal reported that Trump has framed AI safety concerns as an effort that could damage the industry and help China. His posts show his stated opposition to further restraints, while developers and policymakers continue discussing how, or whether, development should be paced.

This story draws on original reporting from Decrypt.