A senior Pentagon official has offered more details on what triggered the dramatic falling-out between the US military and AI company Anthropic. According to a report by news agency Reuters, Emil Michael, Under Secretary of Defense for Research and Engineering, said that when he reviewed the terms of AI contracts signed under the Biden administration, he was alarmed by what he found. Michael’s description in front of an audience at the American Dynamism Summit in Washington on Tuesday (March 3) paints a picture of a military that felt its hands were being tied in the middle of active operations.“I had a ‘holy, holy cow’ moment. There were things … you couldn’t plan an operation … if it would potentially lead to kinetics (or explosions),” Michael said at the summit. He described that the contract had dozens of restrictions baked in and they covered commands responsible for air operations over Iran, China and South America.
‘Previous contracts had sweeping restrictions’
He explained that the contract contained sweeping restrictions on how AI models could be used – which he thinks were so broad that they threatened to shut down military planning in real time. Moreover, he claimed that the contracts were structured so that if a military operator violated the AI provider’s terms of service, the model could theoretically “just stop in the middle of an operation.”While he did not name the AI company whose contracts he was talking about, at the time of his review, Anthropic’s Claude was the only AI model available to the Defense Department on its classified systems.“What we’re not going to do is let any one company dictate a new set of policies above and beyond what Congress has passed,” Michael said.
Emil Michael describes timeline of events in Anthropic vs Pentagon
When Anthropic did not cede to demands by the Pentagon before the deadline, Emil Michael posted on X the timeline of how things unfolded on Friday (February 27).Timeline of events:Today at 9:04pm. No response yet to my calls or messages to @DarioAmodei.Today at 825pm, @AnthropicAI writes “we have not received direct communication from the Department of War.” Today: 5:14pm SecWar tweets supply chain risk designation. Today: I call Dario’s business partner at 5:02pm asking to speak to Dario because he hasn’t gotten back to me. She is typing while we speak and likely has lawyers in the room with no notification to me (that’s a guess! CA is a two party consent law so am hoping no laws were broken!). Says she “will try to locate him.”Today: I call Dario at 5:01. No Answer. I message Dario asking to talk as well. Today: Dario sends an email with redlines 3:51pmToday: @POTUS sends Truth at 3:47pmYesterday: I email Dario comments at 10:54pm. Does this sound like GOOD FAITH?According to the Reuters report, Michael’s concerns escalated further after a senior executive at an unnamed AI company began asking questions about whether its software had been used in what he described as “one of the most successful military operations in recent memory.”Several reports have claimed that the US government raid in January that captured former Venezuelan President Nicolás Maduro were assisted by Anthropic’s tech. Further, the recent the US is also said to have used Anthropic’s AI in Iran strikes — this came hours after Defense Secretary Pete Hegseth declared Anthropic a “supply-chain risk” to national security.
