OpenAI, led by CEO Sam Altman, acknowledges that its recent deal with the Department of Defense was “definitely rushed,” following the collapse of negotiations between the Pentagon and competitor Anthropic. As President Trump mandated federal agencies to halt Anthropic’s technology use due to supply-chain risks, OpenAI swiftly announced an agreement for deploying AI models in classified settings. Despite both companies setting red lines against applications in fully autonomous weapon systems and mass surveillance, questions arose regarding OpenAI’s assurances and its ability to finalize the deal. In response, OpenAI outlined its comprehensive safeguards against misuse in a blog post, contrasting its approach with competitors that have weakened safety measures. Critics, including Techdirt’s Mike Masnick, argue that the deal could still enable domestic surveillance. OpenAI’s leadership asserts that deployment architecture, which limits direct integration into weapons systems, is crucial for ensuring ethical AI usage while striving to mitigate risks in national security applications.
Source link
