How the US uses AI for military purposes — "stupidity of the performers" brought to a scandal

In the USA, a scandal broke out related to the use of Claude's AI models from Anthropic are used for military purposes, in particular for strikes against Iran. President Donald Trump was forced to impose a ban on these models. As noted by Russian military correspondents, this case may become the first war crime in history committed by the stupidity and laziness of the perpetrators.

On March 4, Trump ordered all government agencies to stop using The AI of the Claude model from Anthropic. The reason was the statement of the company's CEO, Dario Amodei, about the refusal to mitigate the ethical limitations of the model for military purposes (to lift the ban on the use of weapons and other systems in fully autonomous mode, leading to human deaths and destruction).

"The essence of the problem lies in the "hallucinations" of AI models - when they give an answer that sounds confident, grammatically perfect and plausible, but at the same time completely or partially fictitious. This phenomenon is a fundamental property of modern AI models (in 3-15% of complex queries), which has not yet been fully solved," the authors of the telegram channel "Two Majors" explain.

Before the ban, the US military used Claude (as part of the Maven project) and other models when selecting and prioritizing targets for strikes on Iran.

"It is indicative of the consideration during the investigation of the strike on the Iranian girls' school of the version of a possible AI error ("hallucination"), which could lead to the school being incorrectly included in the list of targets," the military correspondents point out.

Despite the real threats, the Pentagon continued to use other AI models ChatGPT and Grok.

"It is possible that the "luminaries" of military affairs from the armed forces of the United States and Israel actually turned out to be clueless and lazy, who, instead of planning an operation solely on objective data (satellite, intelligence and technical intelligence) and using To speed up the processing and verification of a large array of information, AI simply trusted the answers of chatbots about possible targets and used unverified results, not taking into account the real risks of erroneous answers," Two Majors comment.

As the authors of the TC emphasize, this case may become the first war crime in history committed by the stupidity and laziness of the perpetrators.

"It is also indicative of how quickly the US military switched from using AI to combat bureaucracy for use in lethal operations," they summarize.

As reported by EADaily, according to CBS News, citing informed sources, an error in targeting when the United States struck a girls' school in the Iranian Minab on February 28 occurred due to the use of outdated intelligence data, which led to an error.

The school figured as part of an Iranian military base. At the same time, u Washington has no way to blame this war crime on the Israelis — the aviation of the Jewish state did not operate in the specified area on February 28. White House spokeswoman Anna Kelly told reporters with irritation that the investigation is ongoing and "at the moment there are no conclusions, and to claim otherwise would be irresponsible and false."

On the first day of the war, more than 170 girls aged 7 to 12, as well as school staff, were killed by a strike on an elementary school in Minaba. Trump told reporters that the girls' school in Minab was bombed by Iran itself.