<\/p>\n
The acrimonious breakup<\/span> between Anthropic and the Department of Defense played out largely in public. But despite the high-profile salvos between the artificial intelligence giant and the Pentagon, ostensibly over the ethics of 21st-century warfare and corporate power,\u00a0their actual falling out hinged on a government contract that has remained hidden away. Through a Freedom of Information Act lawsuit conducted with the help of Legal Advocates for Safe Science and Technology, The Intercept obtained more than 400 pages of documents related to the U.S. military\u2019s contracts with Anthropic and its chief competitors. They illuminate the surprisingly intimate relationship between Big Tech and the armed forces, and reveal new insights about the deals that led to the Anthropic crisis earlier this year.<\/p>\n The documents, many heavily redacted, consist of contract paperwork from deals originally signed in July 2025 between the Department of Defense and Anthropic, Google, OpenAI, and xAI<\/a>, as well as later amendments to those agreements. Through the contracts, each worth up to $200 million, the companies agreed to build prototypes of artificial intelligence tools to \u201cimprove military advantage, military utility, or enhance military decision making\u201d across the whole of the armed forces. These prototypes, aimed at applications from \u201cwarfighting\u201d and automated decision-making to logistics and military intelligence, would go on to form the basis of the artificial intelligence tools now deployed across the Department of Defense\u2019s classified networks. The document set includes one contract<\/a> reflecting OpenAI\u2019s classified network deployment; The Intercept is awaiting the government\u2019s production of equivalent documents from the other contractors.<\/p>\n The contract materials show the close and complex partnership between the Pentagon and the AI industry, and detail the extent to which AI firms themselves are expected to influence how their technology is used to augment the U.S. capacity to wage war.<\/p>\n \u201cThis appears to be the first time the extent of collaboration between frontier AI labs and the Pentagon is spelled out in explicit terms,\u201d Sophia Goodfriend, a University of Cambridge research fellow and non-resident fellow at the Harvard Kennedy School\u2019s Middle East Initiative studying the impact of big data and machine learning on military conflict, told The Intercept. \u201cWhat\u2019s particularly notable is the terms by which engineers are working in lockstep with the department of war to engineer AI systems for surveillance, targeting, and killing.\u201d<\/p>\n Official announcements following the 2025 deals were light on details and heavy on platitudes. \u201cWe must equip our warfighters with 21st-century technology in order to defeat 21st-century threats,\u201d explained<\/a> Pentagon spokesperson Sean Parnell at the time. <\/p>\n Google was similarly<\/a> cagey: \u201cThese advanced AI solutions will enable the DoD to effectively address defense challenges and scale the adoption of agentic AI across enterprise systems to drive innovation and efficiency with agile, proven technology,\u201d the company said.<\/p>\n \u201cWhat\u2019s particularly notable is the terms by which engineers are working in lockstep with the department of war to engineer AI systems for surveillance, targeting, and killing.\u201d<\/p>\n<\/blockquote>\n<\/figure>\n The contract documents for the prototyping project paint a clearer picture, revealing deals that involve far more than a simple sale of software.<\/p>\n An \u201cinformation sharing\u201d provision in all four agreements, for instance, creates a two-way exchange of data and expertise. The companies agreed not only to provide access to their best large language models, but also to \u201cinform DOD and CDAO [Chief Digital and Artificial Intelligence Office] strategy on frontier AI and AI,\u201d as phrased in the agreement with Google.<\/p>\n Under the agreements, the AI labs were approved to receive an array of sensitive information, which could include \u201cbenchmarks datasets for DoD use cases, appropriate briefings on DoD operational missions and threats to inform development \/ release of technologies, their application to DoD operational problems, or DoD plans and strategies for future AI adoption.\u201d<\/p>\n All four companies would in turn provide the Pentagon \u201cfeedback on DoD strategies for frontier AI adoption to ensure that AI capabilities can be delivered and scaled to meet the warfighter\u2019s needs.\u201d The companies also pledged to offer briefings on corporate inner workings, including \u201cfrontier AI model performance, case studies on current or likely future frontier AI applications, or projections of future trends in frontier AI maturation.\u201d<\/p>\n The labs would sometimes be the ones offering intelligence briefings on foreign actors, with Google tapped to inform the Pentagon on \u201cAI tactics and techniques of adversaries\u201d of the United States. These briefings would be conducted in both unclassified and classified settings.<\/p>\n The collaboration also entails Big Tech providing training and educational seminars directly to the military, including \u201cPresentations and briefings at technical exchange meetings on relevant topics, including frontier AI model performance and case studies of frontier AI applications.\u201d One document notes that Google, for example, would lead multiple seminars on \u201cresponsible AI\u201d deployment.\u201d The documents state engineers and policy experts from all four contractors will conduct \u201ctabletop exercises,\u201d gamified simulations<\/a> of real-world scenarios<\/a>, with the Pentagon.<\/p>\n This back and forth between the military and private sector would, the Pentagon hoped, allow for \u201citerative refinement of frontier AI models to address real-world challenges in areas such as intelligence analysis, cybersecurity, and autonomous systems,\u201d according to contract language with all four companies.<\/p>\n The AI companies were contractually scheduled to provide \u201crisk forecasting, and threat ideation exercises\u201d \u2014 predictions of pitfalls and dangers their own products might pose in the future. \u201cFrontier AI labs have advanced risk forecasting tradecraft and a clear vision of the next wave of frontier AI technological developments,\u201d the contracts read. \u201cFrontier AI companies can help DoD gain a better understanding of the risks that might be associated with subsequent near-term frontier AI models or features, to avoid strategic surprise and to mitigate the associated risks.\u201d<\/p>\n <\/p>\n Heidy Khlaaf, chief scientist at the AI Now Institute and former safety engineer at OpenAI, told The Intercept this is a dubious claim with dangerous implications. Pointing to recent disclosures by OpenAI and Anthropic that semi-autonomous large language models broke into computer networks owned by other companies<\/a> during testing, Khlaaf questioned their fitness to help build guardrails in the first place. \u201cThis is a very concerning development,\u201d she said.<\/p>\n According to Khlaaf, the \u201crisk forecasts\u201d so far offered by frontier AI labs have been light on evidence and tend to<\/a> \u201cemphasize self-beneficial skewed risks such as a purported AI arms race and speculative \u2018existential\u2019 risks,\u201d like self-aggrandizing science fiction-style \u201cTerminator\u201d scenarios<\/a>. <\/p>\n Trusting companies to self-report the risks of their own products constitutes a conflict of interest and a \u201csubversion of democratic processes.\u201d<\/p>\n<\/blockquote>\n<\/figure>\n The public would be better served, she said, with a reliance on independent assessment of the risks these companies present, not the word of the companies themselves. Trusting companies to self-report the risks of their own products constitutes both a conflict of interest, Khlaaf added, and a \u201csubversion of democratic processes when AI labs are allowed to take over the arbitration of risk determinations with life-or-death consequences.\u201d<\/p>\n As the prototyping effort progressed, the contract documents show the scope of the project expanded. New versions of the contract show additional deliverables included in the list of what the Pentagon was purchasing. The government redacted descriptions of these new deliverables on the basis of military secrecy and corporate confidentiality.<\/p>\n It was during the negotiation of these subsequent amendments when Anthropic and the Pentagon had their falling out. Anthropic refused to sign a follow-up deal to allow deployment of its technology on classified military networks without contractual prohibitions against use for domestic spying and autonomous weapons, leading Secretary of Defense Pete Hegseth to designate<\/a> the firm a \u201csupply chain risk\u201d in March and ban its services from government use, a decision overturned<\/a> by a federal judge last month.<\/p>\n The Department of Defense quickly reached follow-up agreements with Anthropic\u2019s main competitors \u2014 Google, OpenAI<\/a>, and xAI \u2014 to apply the tools they had been developing since the summer of 2025 to classified military networks. OpenAI\u2019s deal, described<\/a> by CEO Sam Altman as \u201cdefinitely rushed,\u201d was reportedly finalized on February 27. The following day, the company published a blog post<\/a> defending \u201cOur agreement with the Department of War\u201d and insisting there were clauses prohibiting its technology\u2019s use for domestic surveillance.<\/p>\n An amended contract document obtained by The Intercept through its lawsuit dated February 27 contains new language, including a clause noting that OpenAI will embed company engineers within the military to aid U.S. Government employees in implanting and utilizing its AI system. These workers, the document notes, \u201ccan be deployed to warfighting support settings including, but not limited to, combatant commands, service components, and theatre components.\u201d The February 27 amendments also include a section on \u201cSystem Oversight\u201d is redacted in its entirety.<\/p>\n None of the documents released to The Intercept pertaining to xAI or Google\u2019s equivalent agreements contain any such oversight section.<\/p>\n\n
\n