Senate Democrats Demand Probe of AI Targeting Errors in Military

Three senior Democratic senators are demanding federal watchdogs investigate a series of artificial intelligence failures in U.S. military targeting systems, including one incident that sources say nearly triggered armed conflict with China.

Sens. Mark Warner, Jack Reed, and Chris Coons sent a letter Friday to Defense Secretary Pete Hegseth and Director of National Intelligence Jay Clayton calling for immediate inspector general reviews of AI-related targeting errors. The letter specifically references recent reporting that detailed how an AI-generated intelligence report falsely identified cargo aboard a Chinese ship in the Middle East.

The Chinese vessel incident

According to sources familiar with the episode, a special operations command analyst used AI to produce an intelligence assessment that incorrectly concluded a Chinese cargo vessel was transporting nuclear weapons program components. U.S. military officials began preparing an intercept operation before discovering—just before execution—that the AI-generated report was "entirely false."

One source characterized the incident as having "almost started a war." The type of AI "hallucination" involved has occurred multiple times across the intelligence community since agencies began deploying these tools more widely, according to sources.

Pattern of targeting failures

The senators' letter also points to a February U.S. strike on a school in Minab, Iran, that killed nearly 200 people, including children. Three sources previously told reporters that senior military commanders proceeded with that strike despite warnings from critical databases—including an AI-powered system—that intelligence on potential Iranian targets was severely outdated.

"Public reports that AI targeting platforms have repeatedly generated spurious outputs (based on outdated input data or outright hallucination) risk serious operational error, in addition to undermining the confidence of both the American public and international partners," the senators wrote.

Why it matters

These incidents expose a fundamental tension in military AI deployment: the pressure to maintain technological advantage against adversaries like China versus the catastrophic risks of acting on flawed machine-generated intelligence. When AI systems hallucinate in targeting contexts, the consequences extend far beyond software bugs—they can mean unnecessary deaths or unintended wars. The senators' demand for transparency also reflects growing concern that the Pentagon has prioritized speed of AI adoption over governance structures that could catch dangerous errors before they reach operational commanders.

Broader policy clash

The letter arrives amid intensifying congressional debate over AI regulation. Democratic lawmakers have pushed for stronger safeguards, while Republicans including President Donald Trump and House Speaker Mike Johnson have argued that excessive caution could cede ground to China in AI development.

At the Pentagon, Defense Secretary Hegseth has championed broad AI integration into military operations. That effort has encountered resistance, including from AI company Anthropic, which refused to remove safety guardrails that would enable the military to use its Claude model for autonomous weapons and mass surveillance. Anthropic stated its models were not reliable enough for such applications.

The senators are demanding that Hegseth and Clayton provide inspectors general with unrestricted access to investigate both publicly reported incidents and any additional unreported cases of AI targeting errors.

These details were first reported by CNN.