OpenAI Scrambles to Add Surveillance Guardrails to DOD Deal

OpenAI is back at the negotiating table with the Department of Defense, working to bolt on additional protections against domestic mass surveillance as part of their AI partnership. The talks come as Sam Altman's company rushes to implement a deal that was hastily announced last Friday.

The core issue: making sure OpenAI's AI tools don't get weaponized for spying on Americans. The safeguards under negotiation are specifically designed to draw hard lines around how the DOD can deploy the technology domestically.

The scramble to add protections after the announcement — rather than before — raises questions about how buttoned-up the original agreement actually was. OpenAI is now playing catch-up, trying to retrofit privacy guardrails onto a deal that apparently moved faster than its safety framework could keep pace with.