Anthropic's Dario Amodei slams OpenAI's Pentagon contract as 'safety threat'
Anthropic's CEO, Dario Amodei, isn't holding back—he called OpenAI's recent Pentagon contract a "threat to safety" and accused Sam Altman of "straight up lies" about the deal.
Amodei says the real difference is that while OpenAI tries to keep employees happy, Anthropic actually pushes to prevent abuses.
In his words: he said he despised this.
Pentagon canceled $200 million deal with Anthropic
Anthropic rejected certain contractual language in a $200 million Pentagon contract after refusing requests for open-ended access to its AI model, Claude—including use for mass surveillance and autonomous weapons.
As a result, the government subsequently canceled the deal and flagged Anthropic as a supply-chain risk.
Amodei took a jab at OpenAI's PR tactics
Amodei said Anthropic "cannot in good conscience accede" to these government demands because legal protections are weak.
He also took a jab at OpenAI's PR tactics and reports said that ChatGPT saw a 295% spike in uninstalls after news of the deal broke.