How AI safety took a backseat to military money

September 25, 2025·43 min·4 clips

OpenAI removed its ban on military use and signed a $200 million defense deal, marking a major shift in AI priorities.

Host Hayden Field interviews Heidi Klaff, an AI safety expert, about how major AI companies like OpenAI and Anthropic are shifting from ethical safety frameworks to lucrative military contracts. They discuss the concept of 'safety revisionism,' where companies redefine safety standards to accelerate deployment in defense, potentially compromising security and democratic oversight.

As heard by us

A clear look at how AI safety rhetoric collides with military contracts and autonomous weapons risk.

Decoder turns a familiar AI safety debate toward the defense market, with Hayden Field speaking to Heidi Klaff of the AI Now Institute about why leading labs are moving into military applications.

Read the full review in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Decoder with Nilay Patel