The United States military almost started a war with China based on a faulty intelligence report ... which turned out to be completely bogus because A.I. screwed up big time.

Here's the deal ... according to a new report from CNN, an intelligence report claiming a Chinese ship in the Middle East was carrying the components to build a nuclear weapon began to circulate earlier this year amid the ongoing tensions with Iran.

Military planes took to the air and armed military personnel were ready to board the ship ... when someone took a closer look at the report and realized a huge mistake had been made.

The report was "entirely false" according to one CNN source ... because a special operations command analyst had used A.I. to generate the report and the program had misidentified the ship's cargo. It's unclear what the ship was actually carrying, but it wasn't nuclear bomb parts.

The chatbot had apparently been looking into intelligence gathered about the ship's manifest ... and used "open-source intelligence with secret signals intelligence in government holdings" to come to its boneheaded conclusion.

The U.S. government isn't shying away from A.I. ... in fact, Secretary of War Pete Hegseth has pushed a new initiative -- his department's "Artificial Intelligence Acceleration Strategy" -- to use these tools more regularly for defense.

This reliance on robots will certainly draw more scrutiny, as many in the A.I. field warn against the rapid advancement of the technology... with some fearing it could end up killing all of humanity.

Bottom line ... A.I. sounds like it nearly started WW3.