Anthropic CEO calls OpenAI's messaging around military deal 'straight up lies'

Source: Hacker News

Memo from Anthropic CEO

Anthropic co‑founder and CEO Dario Amodei criticized OpenAI chief Sam Altman in an internal memo. Referring to OpenAI’s dealings with the Department of Defense (DoD), Amodei called the arrangement “safety theater” and accused OpenAI of “straight up lies” for presenting Altman as a “peacemaker and dealmaker.”

“The main reason [OpenAI] accepted [the DoD’s deal] and we did not is that they cared about placating employees, and we actually cared about preventing abuses,” Amodei wrote.

DoD Deal with OpenAI vs. Anthropic

• OpenAI announced a new defense contract that includes technical safeguards. Altman stated the contract would protect against the same red lines Anthropic had asserted. Source: Sam Altman’s X post and the related TechCrunch article.
• Anthropic and the DoD failed to reach an agreement after the military requested unrestricted access to Anthropic’s technology. Anthropic, which already held a $200 million contract with the DoD, insisted the department affirm it would not use the AI for domestic mass surveillance or autonomous weaponry. Source: TechCrunch.

OpenAI’s blog post explains its stance:

“Our contract allows use of our AI systems for ‘all lawful purposes.’ It was clear in our interaction that the DoW considers mass domestic surveillance illegal and was not planning to use it for this purpose. We ensured that the fact that it is not covered under lawful use was made explicit in our contract.” OpenAI blog

Anthropic’s Conditions

Anthropic specifically objected to the DoD’s insistence on “any lawful use” of its AI, arguing that such language could permit activities the company deems abusive, such as:

• Domestic mass surveillance
• Autonomous weaponry

Anthropic’s position was outlined in its public statement to the Department of War. Statement

Public Reaction

• ChatGPT uninstall rates surged by 295 % after OpenAI’s DoD deal was announced. TechCrunch report
• Amodei noted that the public and media largely view OpenAI’s deal as “sketchy or suspicious,” while Anthropic is seen as a “hero.” He added that the narrative is resonating with most users, though it may still affect some OpenAI employees.

“I think this attempted spin/gaslighting is not working very well on the general public or the media… It is working on some Twitter morons, which doesn’t matter, but my main worry is how to make sure it doesn’t work on OpenAI employees.” – Dario Amodei

Context

Critics have highlighted that legal definitions of “lawful use” can evolve, meaning practices considered illegal today might become permissible in the future. This uncertainty fuels ongoing debate over AI governance and military applications.