US threatens Anthropic with deadline in dispute on AI safeguards

The AI developer laid out red lines on military use of its products, a source said.

Anthropic has dropped a central pledge from its 2023 Responsible Scaling Policy, which previously barred it from training AI systems unless it could guarantee adequate safety measures, amid rapid AI advancements and intensifying industry and geopolitical competition.

Defense Secretary Pete Hegseth has given San Francisco artificial intelligence company Anthropic until Friday to roll back safety restrictions on its AI model, Claude, for military use, or risk losing a $200 million Pentagon contract and being designated a 'supply chain risk'.