They say OpenAI got “same deal” Anthropic was banned for. They’re not the same.

OpenAI's Pentagon deal differs from the one Anthropic was banned for, despite similar wording. Anthropic asked for no fully autonomous weapons without human oversight, meaning a human involved in the decision, while OpenAI's deal refers to "human responsibility for the use of force," which can be met after the fact. On surveillance, Dario says current law has not caught up with AI, which could let the government assemble bought movement and browsing data into a complete picture of a person's life without breaking any law. OpenAI's deal says the Pentagon "reflects them in law and policy," which relies on the existing law Anthropic considered insufficient.