OpenAI will build AI for the US military
OpenAI just signed a deal with the US Department of Defense to bring its AI tech onto classified government networks.
But this isn't a free-for-all—the agreement bans mass surveillance at home and says humans must stay in charge when it comes to autonomous weapons.
OpenAI will build special safety systems
OpenAI will build special safety systems and send its own engineers to make sure everything follows the rules.
Their AI will only run on cloud servers, so it can't be used for fully autonomous weapons or spying on Americans.
Plus, OpenAI has kept the right to refuse requests that would require it to act outside legal limits.
Altman wants all AI companies to have the same standards
CEO Sam Altman wants these safety standards applied to all AI companies working with the government—not just OpenAI.
He shared that they'd actually turned down similar deals before (and Pentagon officials said they were surprised OpenAI was willing to consider classified work), but now they're stepping up—if everyone plays by the same fair rules.