OpenAI-US Defense Cooperation Formalized
''No Domestic Surveillance or Autonomous Weapons Without Human Control'' Codified

OpenAI CEO Sam Altman announced agreement with the US "Department of War" (his symbolic term for the Department of Defense) to deploy AI models on classified networks. Three core conditions codified in the contract: (1) No use for large-scale domestic surveillance — reflecting constitutional privacy protection principles; (2) Human responsibility for use of force — all autonomous weapons operations require human final judgment and accountability; (3) Technical safety measures for malfunction/misuse prevention — deployment only on cloud networks (not on-premise), maintaining central control, real-time patching, and remote shutdown capability. The "Department of War" framing: Altman''s use of this symbolic term signals that AI military applications have moved from abstract policy discussion to operational reality — while simultaneously pre-empting internal opposition through explicit safety commitments. Internal industry tension: hundreds of OpenAI and Google employees signed open letters expressing concern about government overreach potential; previous AI-military contracts (Google Maven) led to significant employee protests. Two key controversy areas: (1) Domestic surveillance — AI analyzing large-scale visual, audio, and text data to track citizen activities could conflict with constitutional privacy protections; (2) Autonomous lethal systems — UN-level autonomous weapons regulation discussions are ongoing; how "human control" is actually enforced in operational environments is the critical variable. Altman stated he asked for the same safety conditions to be applied to other AI companies in military contracts — an attempt to standardize minimum ethical requirements across the industry rather than creating competitive disadvantage for OpenAI through its constraints. Outstanding questions: disclosure scope of contract details; definition of "surveillance"; enforcement of autonomous weapons control in practice; application to overseas operations.