Anthropic relaxes its AI safety policy to match rivals

Claude maker Anthropic announced it is scaling back its safety policy to stay competitive in the AI race. The decision follows a tense meeting between CEO Dario Amodei and Defense Secretary Pete Hegseth.

Axios reports that Anthropic was given a Friday deadline to lower safeguards on the military’s access to its AI model or risk losing government contracts. It currently has a $200 million defense deal with the Pentagon.

The AI company was at a crossroads with the Defense Department on enabling technology that would allow mass surveillance of Americans and the development of autonomous weapons that fire without human intervention.

Anthropic’s shift marks a dramatic departure from Amodei’s vision of it as a safety-first firm. Former OpenAI executives founded the company specifically to prioritize safety over speed.

Despite the retreat, Anthropic maintains it remains committed to its core safety policy and to regularly publishing third-party risk reports related to its future AI models.