Gemini hacked three companies in May during a test by Irregular; Google says the model stopped after determining it had accessed real companies' systems (Wall Street Journal)

During a May test by Irregular, Gemini reportedly "hacked three companies," an incident Google claims was halted by the model itself after it recognized it had accessed real company systems. This event, while similar to other AI model hacks, was not considered an instance of "model misalignment" by Google. The details were reported by the Wall Street Journal.