President Donald Trump has directed all federal agencies to immediately stop using technology from Anthropic PBC. The ban extends to military suppliers, as announced by Defense Secretary Pete Hegseth. These actions stem from disagreements over the safety guardrails built into Anthropic's AI models.
The Pentagon previously awarded Anthropic a $200 million AI contract, intending to explore national security applications. However, clashes arose over Anthropic's policy prohibiting AI use for mass surveillance or autonomous weapons development, which the Pentagon sought to override for "all lawful purposes."
Despite Pentagon offers of concessions, Anthropic stated the proposed compromises still allowed safeguards to be disregarded. Following a meeting with Defense Secretary Hegseth, Anthropic was given an ultimatum to revise its policy. The Pentagon has now designated Anthropic a supply chain risk, mandating the phase-out of its models from government networks within six months.
This development follows similar concerns raised by OpenAI, whose CEO noted the company applies equivalent safety "red lines" to its models. The Pentagon has previously approved xAI's Grok for classified networks, despite safety reservations from federal agencies.