What happened
OpenAI finalised an agreement with the US Department of War on 28 February, establishing specific guardrails for AI system deployment in classified environments. OpenAI states its AI will not independently direct autonomous weapons where human control is required, nor will it be used for unconstrained monitoring of US persons' private information or domestic law-enforcement activities except as permitted by applicable law. The company retains full discretion over its safety stack, involves cleared personnel, and can terminate the contract if terms are violated. OpenAI claims this agreement includes more guardrails than previous classified AI deployments, including Anthropic's.
Why it matters
This agreement establishes a precedent for AI developers negotiating military contracts, defining explicit limitations on autonomous weapon control and surveillance. For founders and legal teams, it clarifies the potential for retaining safety oversight and contractual termination rights in defence partnerships. This contrasts with President Trump's recent directive to terminate federal contracts with Anthropic, which refused to loosen its AI safety standards regarding mass surveillance and autonomous weapons, highlighting diverging approaches to government engagement. Procurement teams must now evaluate vendor agreements against these new benchmarks for ethical AI deployment in sensitive sectors.
Subscribe for Weekly Updates
Stay ahead with our weekly AI and tech briefings, delivered every Tuesday.
