The Pentagon's Private-Sector A-Team
Defense Secretary Pete Hegseth has assembled a controversial team to lead the Pentagon's AI negotiations, including former Uber executive Emil Michael and private equity billionaire Steve Feinberg. This comes as the Department of Defense engages in hardball contract renegotiations with Anthropic over its AI use policies.
Key Figures in the Pentagon's AI Push
Emil Michael (Pentagon CTO) Michael, spearheading negotiations with Anthropic, was formerly second-in-command at Uber during Travis Kalanick's tenure. He was pushed out in 2017 after investigations revealed he perpetuated a culture of sexual harassment. Michael previously suggested hiring opposition researchers to dig up dirt on journalists critical of Uber and was involved in controversies around Uber's 'God Mode' tool used to track users.
Steve Feinberg (Deputy Secretary) Founder of Cerberus Capital Management ($65B in assets), Feinberg has been blamed for the death of Chrysler as an auto manufacturer. During his 2025 Senate confirmation, Democrats raised concerns about conflicts of interest due to Cerberus' investments in defense contractors like DynCorp. In 2023, Cerberus launched a venture capital arm investing in early-stage national security companies.
Sean Parnell (Chief Spokesperson) An Army veteran who attempted a 2021 Senate run in Pennsylvania with Trump's endorsement, Parnell was forced to drop out after his ex-wife made allegations of serious physical and psychological abuse during a custody hearing.
The Single-Supplier Vulnerability
The Pentagon faces a critical dilemma. A 2024 Biden administration national security memorandum requires the DoD to maintain contracts with at least two frontier AI labs cleared for classified information to prevent single-supplier vulnerability.
Currently, only Anthropic's Claude is cleared for classified use. The Pentagon suddenly granted xAI's Grok access to classified systems—even though defense insiders consider Grok less advanced and reliable. Google is reportedly close to signing a deal for Gemini, while OpenAI believes ChatGPT needs improved safety features before classified deployment.
The Anthropic Standoff
The Pentagon continues to designate Anthropic as a 'supply-chain risk' if the company doesn't comply with demands about its acceptable use policy. Anthropic has reportedly refused to allow its AI to be used for autonomously killing people without human input—a stance that has created tension with defense officials.
As one Defense official told Axios: 'The only reason we're still talking to these people [Anthropic] is we need them and we need them now. The problem for these guys is they are that good.'
Source: The Verge
