Anthropic CEO attacks OpenAI's Pentagon deal as "safety theater" while investors scramble for de-escalation

Editor's take

Anthropic's CEO publicly criticized OpenAI's Department of Defense contract, labeling it largely performative safety measures and alleging political retribution from the Trump administration.

This highlights a growing tension within the AI safety discourse, particularly between major players like OpenAI and Anthropic. The framing of the Pentagon deal as "safety theater" by Amodei, whose company has also pursued government contracts, suggests a strategic divergence in how AI companies are navigating national security interests and public perception, potentially influencing future defense procurement and regulatory approaches.

Future scrutiny should focus on whether OpenAI can demonstrably integrate its AI models into military operations with verifiable safety protocols, and if other AI firms will adopt similar stances to differentiate themselves. The long-term impact on AI development's intersection with defense will depend on the concrete outcomes of this public disagreement and any subsequent regulatory or contractual adjustments.