Pentagon Reportedly Used Anthropic Claude in Middle East Strike Despite Trump Federal Ban
A bombshell report from The Wall Street Journal, citing sources familiar with the matter, reveals that the U.S. Department of Defense (Pentagon) utilized Anthropic’s Claude AI during a recent joint military operation with Israel against Iran. This revelation comes at a particularly sensitive time just hours after President Donald Trump issued a direct executive order for all federal agencies to cease using the AI platform.
The Role of AI in Combat Operations
According to the report, Claude AI was integrated into the mission’s aerial operations to provide high-level strategic support. Its primary functions included:
Intelligence Assessment: Processing vast amounts of reconnaissance data in real-time.
Target Identification: Assisting in pinpointing high-value military objectives.
Tactical Simulations: Modeling various scenarios to support command decisions under pressure.
The Transition Dilemma
This operational use highlights the Pentagon’s deep reliance on Claude, which has apparently become a cornerstone of the U.S. military’s AI infrastructure. Military analysts suggest that while the administration is pushing for a shift toward other AI providers, the transition period will be fraught with challenges. Completely decoupling such a deeply integrated system without compromising operational readiness could take months, if not years.
This incident reflects the phenomenon known as "Shadow AI" within security agencies: technology becoming so deeply ingrained in operating systems that frontline personnel prioritize "following the mission" over "following policy," because actual combat outcomes (Mission Success) outweigh political orders at the time.
Analysts believe the Pentagon's continued use of Claude, despite conflicts with the company, is due to its vast "context window" capabilities (the ability to read and process thick intelligence documents simultaneously) and low hallucination accuracy, crucial for identifying targets that may inflict collateral damage.
Even though OpenAI has signed a new contract with the Pentagon, transitioning AI systems for real-world combat isn't simply a matter of "changing apps." It requires fine-tuning of classified data and rigorous security testing (stress testing). A WSJ report indicates that OpenAI's technology may not be as readily ready for battlefield use as Claude.
From Creative Tool to War Machine? Why Millions are Quitting ChatGPT This Week.
Source: The Wall Street Journal
Comments
Post a Comment