The US almost boards a Chinese ship after an artificial intelligence error

The United States prepared an operation to intercept and board a Chinese ship in the Middle East after a report prepared with artificial intelligence (AI) wrongly indicated its cargo as components for a nuclear weapons program, according to CNN reported this Friday.

The incident occurred at the beginning of the year, in the context of the war with Iran, according to four sources with knowledge of the case cited by the American media.

Two of them indicated that armed soldiers were getting ready to board the boat, while several US aircraft flew over the area.

The operation was suspended shortly before it was to be carried out, when US officials more closely reviewed the information and determined that the chatbot that an analyst had used incorrectly identified the material being transported.

According to CNN, the tool combined information from open sources with classified intelligence stored by the US government and wrongly concluded that the payload was linked to a nuclear weapons program. "The internal tools are, for the most part, simple copies of the commercial ones with a facelift," a former senior US official familiar with the artificial intelligence systems used by military and intelligence analysts told the media.

The chain could not determine what the ship's actual load was or whether the tool was a commercial product or a system developed by the US Government.

One of the sources described the report as "completely false" and stated that a possible intervention against a Chinese vessel could have caused an escalation between Washington and Beijing. "A war was about to start," says the source.

The case comes at a time when the use of AI tools by the US Army and intelligence services is increasing, as is criticism of this technology.

Last weekend, Anthropic CEO Dario Amodei published an essay in which he called for slowing the pace of industry development in the face of growing security risks from AI.

The text joins statements along the same lines from Sam Altman, executive director of OpenAI, and billionaire Elon Musk.