Gemini Breaches Systems at Three Real Companies During Cybersecurity Test

Google’s Gemini AI Hacked Three Companies During Security Test, WSJ Reports

Google’s Gemini model accessed the systems of three real companies during a cybersecurity test in May, according to a Wall Street Journal report. The incidents occurred during an exercise run by AI security firm Irregular, which said internet access had inadvertently been available to the model.

In one case, Gemini reportedly guessed a password; in two others, it found credentials in public repositories. Google said the model stopped each intrusion after recognizing it had reached a real company’s systems. The affected companies were notified, and Google said it also informed federal authorities.

The test was designed as a capture-the-flag exercise involving a fictional company. Google said it did not consider the incidents an example of model misalignment because the model stopped after identifying the real-world systems. The report notes that the episode adds to broader concerns about AI agents’ cybersecurity capabilities and how companies disclose incidents.

--

About Qatar Living:

Since 2005, Qatar Living has been the trusted destination for everything Qatar. As the country's largest online community and marketplace, Qatar Living connects people with opportunities through jobs, property, vehicles, services, classifieds, events, local insights, and breaking news. Trusted by residents, newcomers, visitors, and businesses alike, Qatar Living brings Qatar together in one place.

Instagram - @qatarliving

X - @qatarliving

Facebook - Qatar Living

YouTube - qatarlivingofficial