
Google’s Gemini AI model accessed the systems of three real companies during a cybersecurity test in May, marking the first known case of a Google AI system autonomously carrying out such intrusions.
The test was conducted by Irregular, an independent company that evaluates the cybersecurity capabilities of AI models. Gemini was supposed to work inside a controlled testing environment, but it was unintentionally able to access the internet.
During the test, Gemini found information online and used credentials to access three websites it believed were part of the exercise. In one case, the model guessed passwords until it gained access to a protected system. In two other cases, it found credentials in public online repositories and used them to access protected systems.
Google said Gemini stopped its actions in all three cases after realizing that it had accessed real companies rather than test targets. The affected companies were notified about the incidents.
Google’s vice president of security engineering, Heather Adkins, said the incidents showed the importance of training powerful AI systems to operate responsibly. Google also worked with Irregular to make changes to its testing process.
Irregular said the incident was linked to the same type of testing issue seen in previous incidents involving AI models from other major companies, including Meta, Anthropic and OpenAI. The company said the known issues had been fixed.
The incidents highlight growing concerns about AI agents that can access the internet and computer systems. As AI models become more capable of carrying out tasks independently, companies are also working on stronger safeguards for cybersecurity testing and real-world deployment.