Google’s artificial intelligence model gained access to the Internet and penetrated the secure systems of 3 companies in what is being described as the first known example of a cyber attack by Google’s artificial intelligence system autonomously.

According to the Wall Street Journal, the incidents occurred in May during a cybersecurity audit conducted by an independent cybersecurity firm.

According to the report, in one incident, an artificial intelligence model gained access to a secure system by repeatedly guessing passwords, while in the other 2 incidents, Gemnai obtained confidential information from public data through which it accessed the secure systems.

A spokesperson for the agency said the issue was similar to other AI companies and that all agencies were notified in late July.

A Google Gemini spokesperson says that all known issues have been resolved by us several weeks ago.

It should be remembered that earlier such incidents have also been reported by Meta, Anthropic and OpenAI.

These incidents have raised important questions about security measures given to artificial intelligence systems with greater autonomy and access to the Internet and computer systems.