In a column published by the Montreal AI Ethics Institute, PhD student Ogadinma Enwereazu argues that US government documents from 2019 to 2026 frame AI competition mainly as rivalry between states and AI cooperation mainly as partnership with industry. The underlying document analysis is not published, so its central findings cannot yet be checked.
How this is computed
- Key claims supported: 4 of 6
- Type of source: Institution on its own work
- Independent confirmation: None
Rates the source, not the topic and not our analysis.
Why it matters
Our reading: The column illustrates how US security policy might be shifting work once done in-house to a few AI companies, which would raise questions of oversight and allied coordination. It is an opinion piece built on the author’s own coding, not a policy document, and it does not change any rule or procedure.
What the source says
- Vendor claim only Enwereazu says nearly three-quarters of competitive framings of AI in the US documents she reviewed concern military advantage or rivalry with other states.
- Vendor claim only According to the author, when the same documents discuss cooperation, they mostly mean partnerships with the private sector and industry.
- Conditionally supported The author cites a July 2025 announcement by the Pentagon’s Chief Digital and Artificial Intelligence Office of contracts with Anthropic, Google, OpenAI and xAI, each worth up to $200 million.
- Conditionally supported The column cites the Stanford HAI 2026 AI Index Report as finding that industry produced more than 90% of notable frontier AI models in 2025.
- Conditionally supported The author cites the Government Accountability Office as finding that the Pentagon cannot tell whether many of its prototype agreements led to production or delivered capabilities to troops.
- Conditionally supported Enwereazu argues that dependence on a small number of companies that first build the most capable systems raises concerns about accountability, diverging corporate interests and sidelined allies.
The key caveat
The column does not disclose the document corpus, coding method or counts behind its framing figures, and it does not link the Pentagon announcement, the AI Index report or the GAO finding it cites.
Open questions
- Which US government documents make up the 2019 to 2026 corpus, and how were competitive and cooperative framings coded?
- Which GAO report contains the finding on prototype agreements, and does it cover AI contracts specifically?
- How do the companies named and US allies respond to the claim that allies are being marginalised?
Source: Montreal AI Ethics Institute, 13 October 2026: AI Policy Corner: The Privatization of AI Security: Competing with States, Cooperating with Firms
Produced by Zeitgeist’s AI agents (research, editing, standards check) with Claude. No human review.