Pentagon Halts Use of Anthropic AI Tools Following National Security Blacklist
Key Takeaways
- The US Department of Defence has officially stopped using Anthropic AI products after blacklisting the firm as a supply chain risk.
- Anthropic's Claude was deeply integrated into Palantir's Maven system and used in military intelligence tasks, including operations against Iran.
- The dispute began when Anthropic refused military demands to disable safety guardrails regarding autonomous weapons and mass surveillance.
The US Department of Defence has formally halted all use of Anthropic's artificial intelligence tools, a department official confirmed on Monday. The decision follows a directive by Defence Secretary Pete Hegseth in February designating the AI firm a supply chain risk on national security grounds. Despite the directive, multiple sources familiar with the matter confirmed that Anthropic's Claude model was still actively utilized as recently as last week for military research, analysis, and operational intelligence gathering.
Anthropic’s models were deeply embedded in the Pentagon’s Maven Smart System, a core data and intelligence platform operated by Palantir. According to defence contractors and former officials, Claude was critical for processing satellite imagery, drone footage, and intelligence for operations against Iran and target identification. Former defence official Lauren Kahn noted that removing Claude took months because advanced AI architectures are not simple to extract: "Once they become integrated it can be painful to remove them," Kahn explained.
The standoff began after the Pentagon pressured Anthropic to strip safety guardrails from Claude to grant the military unfettered access to its AI models. Anthropic refused, citing deep concerns over mass surveillance and the deployment of autonomous weapons systems. In response, Hegseth invoked a national security supply-chain risk label typically reserved for adversary foreign firms, prompting Anthropic to file a federal lawsuit against the Trump administration calling the designation unlawful.
While the Pentagon has pivoted to alternative contracts with competitors including Google, xAI, and OpenAI, signs of a political thaw have recently emerged. Former Anthropic CEO Dario Amodei met twice with President Donald Trump at the White House last week to discuss ongoing AI safety concerns. Trump publicly praised Amodei as "fantastic," leaving the future trajectory of Anthropic's government relationship uncertain despite the operational ban.
Subscribe to the Global Briefing
Receive our flagship executive dispatch every morning. Essential geopolitical analysis, financial benchmarks, and breaking world events delivered to your inbox.
Related Stories & Follow-ups
Editorial Wire
TechnologyAnthropic Confirms Houthi Rebels Used Claude AI to Develop Advanced Missile Guidance Systems
TechnologyTesla Unveils Roadster Featuring SpaceX-Derived Thruster Technology for Flight
TechnologyElon Musk Unveils Flying Tesla Roadster Prototype with 2028 Production Target
Reader Discussion
Leave a Comment
0 Comments
Owner
Editorial Oversight & Machine-Assisted Reporting
Human-in-the-Loop VerifiedEditorial Oversight & Machine-Assisted Reporting: This dispatch was synthesized and localized via OPA automated intelligence under continuous human editorial supervision.
Wire Source
This report was synthesized from BBC World under journalistic fair use.