Trump warns AI could be weapon if used ‘improperly’ after blocking Anthropic models

US President Donald Trump said artificial intelligence could transform into a weapon if “improperly used,” following an administration move to disable advanced models from the firm Anthropic. “We have a situation with Anthropic, and we didn’t like what they were doing,” Trump told Axios in an interview released Friday. He said that a “competitor and a part owner” reported the company to federal officials, leading to an export control directive that forced the suspension of the “Fable 5” and “Mythos 5” systems to prevent potential “jailbreaking.”

Compliance and regulation

Trump said Anthropic CEO Dario Amodei acted “responsibly” once the government intervened, warning that non‑compliance carries severe penalties. While describing AI as “bigger than the internet,” he said the benefits far outweigh the risks. Asked if he would use the Defense Production Act to regulate the sector, he said aggressive measures would not be necessary if companies cooperate. When asked if Anthropic or its CEO pose a national security threat, Trump replied: “Well, not now, but a week ago, maybe.”

Anthropic’s response

Anthropic said the US government had issued an export control directive to suspend all access to the models by foreign nationals, extending the restriction even to its own employees. The company said the letter did not provide specific details on the national security concerns, and that the team had reviewed a demonstration of a technique used to identify minor vulnerabilities.