WASHINGTON, DC – JANUARY 29: US Secretary of the Military Pete Hegseth (heart) speaks to the Cupboard as US President Donald Trump (left) and Secretary of Commerce Howard Lutnick (proper) hear within the Cupboard Room on the White Home on January 29, 2026 in Washington, DC. President Trump is in session because the Senate plans to vote on a spending package deal to avert one other authorities shutdown, whereas Democrats look ahead to a deal to contemplate funding the Division of Homeland Safety. (Photograph by Win McNamee/Getty Photos)Getty Photos
Can AI firms set limits on the place and the way the army makes use of their fashions? Anthropic is in heated negotiations with the Pentagon, refusing to adjust to new army contract provisions that require loosening guardrails on AI fashions. That provision would authorize “any lawful use” of Americans, together with mass surveillance and totally autonomous deadly weapons.
Emil Michael, the Pentagon’s chief expertise officer, has known as for Anthropic to be labeled as a “provide chain danger,” usually reserved for nationwide safety threats, if it doesn’t comply. Anthropic’s rivals OpenAI and xAI have reportedly agreed to new phrases, however after assembly with Secretary of Protection Pete Hegseth on the White Home, Anthropic CEO Dario Amodei nonetheless refused to cross his firm’s purple line, saying, “Threats don’t change our place. We can’t in good conscience adjust to their calls for.”
For the most recent info on the conflict between AI firms and the Division of Protection, see right here…
No must have unsupervised killer robots Anthropic rejects new Pentagon circumstances, stands agency on autonomous deadly weapons and mass surveillance Pete Hegseth’s group of Pentagon AI allies consists of former Uber govt and personal fairness billionaire Inside Anthropic’s existential negotiations with the Pentagon
