Anthropic moves toward deal with US to lift curbs on AI models

Anthropic PBC is working to resolve a recent dispute with the Trump administration over export restrictions on its advanced artificial intelligence models. The company has dispatched senior technical staff to Washington, D.C., to engage in discussions with government officials, including Commerce Secretary Howard Lutnick and National Cyber Director Sean Cairncross, in an effort to lift the curbs. The restrictions, imposed by the Trump administration, prohibit accessing Fable 5 and Mythos 5.

The administration’s decision followed research from Amazon that demonstrated certain safeguards on Fable 5 could be bypassed, prompting officials to demand the model’s removal. Anthropic complied by shutting down access to both models globally. The company argues that the restrictions could hinder cybersecurity efforts, as its partners identify and patch software vulnerabilities.

Cybersecurity experts and industry leaders have expressed concerns that the restrictions could harm national security and innovation. Some have criticized the administration’s response as overly cautious, noting that similar vulnerabilities exist in other publicly available AI models. Anthropic, which is preparing for a potential public offering, has emphasized its commitment to safety and transparency.

The ongoing discussions highlight the broader regulatory challenges facing the AI industry as governments seek to balance innovation with security concerns.