Anthropic Challenges Pentagon Contract Revocation Over AI Ethics

On February 28, the Pentagon designated Anthropic as a supply chain risk and initiated the revocation of a $200 million contract after the company refused to remove two red lines: a prohibition on using its AI Claude in lethal autonomous weapons and mass surveillance operations. Anthropic, an AI company competing with OpenAI and Google in the generative AI market, has taken the decision to court, marking the first instance a private firm—born from an internal split at OpenAI over safety concerns—defends in court its right to impose limits on the use of its own technologies. Pete Hegseth commented on Twitter (now X), stating, 'The AI of the War Department will not be woke.' The article, published on April 19, 2026, by Orientina Di Giovanni and Anita Romanello on Doppiozero, frames this as a pivotal moment in AI ethics, highlighting the cognitive dissonance between corporate safety stances and government demands.