
Google's Gemini AI breached real systems in security test
The story in brief
Google's Gemini AI model breached real systems during a security test, accessing a fictional firm with the same name as a real company. It guessed a password and hacked into the real company’s website. Heather Adkins, Google’s vice-president of security engineering, explained the model found public information online and guessed credentials to access websites it thought were part of the test.