Following a strategic agreement, OpenAI’s chief executive officer, Sam Altman, said on Saturday that the US Department of Defence has authorised the deployment of OpenAI’s artificial intelligence models on its classified network.
The company’s operationalisation of its models within safe, classified cloud settings has been concluded, according to Altman, with the Pentagon. Speaking with officials from the defence department, he noted that there was “deep respect for safety” and agreement on the prudent use of AI.
Altman stated in an X post that the deal is in line with OpenAI’s larger goal of increasing access to the benefits of AI while simultaneously ensuring its safety. The two principles he deemed non-negotiable, he said, are the preservation of human accountability in the use of force (including autonomous weapon systems) and the ban of domestic mass surveillance.
These protections, according to Altman, are already part of the agreement and are based on US law and policy from the defence department. To further guarantee that OpenAI’s systems operate as expected in classified environments, he said the company will be implementing extra technical restrictions.
The agreement calls for the corporation to send out field engineers to help with the safe rollout and monitoring of the models. Only within the secure cloud can the systems function.
