OpenAI: Pentagon Deal Features Tight Safeguards for Classified AI
OpenAI announced on Saturday that the agreement it reached a day prior with the Pentagon to implement technology on the US defense department’s classified network incorporates further safeguards to ensure the protection of its use cases. On Friday, US President Donald Trump instructed the government to cease collaboration with Anthropic, while the Pentagon announced plans to classify the startup as a supply-chain risk. This decision represents a significant setback for the artificial intelligence lab following a confrontation regarding technology safeguards. Anthropic stated, “it would challenge any risk designation in court.” Soon after, rival OpenAI, which is supported by Microsoft, Amazon, SoftBank, and others, announced its own deal late on Friday. “We think our agreement has more guardrails than any previous agreement for classified AI deployments, including Anthropic’s,” OpenAI said on Saturday.
The AI firm stated that the contract with the Department of Defense, which has been renamed the Department of War by the Trump administration, establishes three critical boundaries: OpenAI technology is prohibited from being utilized for mass domestic surveillance, directing autonomous weapons systems, or making any high-stakes automated decisions.In our agreement, we safeguard our red lines via a broader, multi-faceted strategy. “We retain full discretion over our safety stack, we deploy via cloud, cleared OpenAI personnel are in the loop, and we have strong contractual protections,” OpenAI stated.
The Pentagon has entered into agreements valued at as much as $200 million each with prominent AI laboratories over the past year, including Anthropic, OpenAI, and Google. The Pentagon aims to maintain complete flexibility in defense strategies, disregarding cautions from the technology’s developers about the risks of utilizing unreliable AI in weaponry. OpenAI has warned that any violation of its contract by the US government could lead to termination; however, it noted, “We don’t expect that to happen.” The company asserted that competitor Anthropic ought not to be classified as a supply-chain risk, emphasizing, “We have made our position on this clear to the government.”
