OpenAI CEO Sam Altman announced an agreement with the Department of Defense to deploy the company’s AI models within a classified government network. Altman highlighted the Department’s emphasis on safety and collaboration for optimal outcomes. According to Fortune, OpenAI will have autonomy to establish a ‘safety stack,’ a comprehensive system of controls ensuring responsible AI use. Notably, the government will respect OpenAI’s ‘red lines,’ prohibiting AI deployment in autonomous weapons, mass surveillance, or critical decision-making scenarios.
This milestone follows public tensions between the Department of Defense and OpenAI’s competitor, Anthropic, resulting in the apparent termination of Anthropic’s defense contracts. OpenAI will retain authority over safeguard implementation, model selection, and deployment locations, confining usage to cloud systems rather than on-field applications like drones.
Source: Tech-Economic Times