Anthropic CEO accuses OpenAI of misleading statements over military AI partnership

Anthropic co-founder and CEO Dario Amodei is clearly, perhaps unsurprisingly, displeased with OpenAI CEO Sam Altman. In an internal memo to employees, as reported by The Information, Amodei described OpenAI’s relationship with the Department of Defence as “safety theatre.”

“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.

Last week, Anthropic and the U.S. Department of Defence (DoD) were unable to reach an agreement onhe military’s demand for unrestricted access to the AI company’s technology. Anthropic, which already held a $200 million military contract, maintained that the DoD needed to formally confirm it would not use the company’s AI systems to support domestic mass surveillance or autonomous weapons.

Instead, the DoD, referred to under the Trump administration as the Department of War, finalised an agreement with OpenAI. Altman said the new defence contract his company signed would include safeguards against the same red lines Anthropic had pushed for.

In a message to staff, Amodei called OpenAI’s public framing “straight up lies,” saying Altman was wrongly “presenting himself as a peacemaker and dealmaker.”

Amodei may not be speaking purely from frustration. Anthropic’s main objection centred on the DoD’s insistence that its AI be made available for “any lawful use.” OpenAI, meanwhile, said in a blog post that its agreement permits the use of its 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,” OpenAI’s blog post said. “We ensured that the fact that it is not covered under lawful use was made explicit in our contract.”

Critics have noted that laws can change over time, and actions considered illegal today could become legally permitted later.

Public opinion also appears to be leaning toward Anthropic. Uninstalls of ChatGPT reportedly surged 295% after OpenAI entered into its agreement with the DoD.

“I think this attempted spin/gaslighting is not working very well on the general public or the media, where people mostly see OpenAI’s deal with the DoW as sketchy or suspicious, and see us as the heroes (we’re #2 in the App Store now!),” Amodei wrote to employees. “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.”