Google’s Gemini AI Model Hacked 3 Real Time Companies During Cybersecurity Test

Google has come forward and disclosed that its AI agents have hacked another company. The tech giant gave confirmation that its Gemini AI models have hacked into three real companies during the cybersecurity test, but the AI stopped on its own after it realised what had happened.

As per a recent report from the Wall Street Journal, Google’s Gemini model accessed the internet and hacked three outside companies during its cybersecurity test run by the Israeli startup irregularly in the month of May. This marks the first known case of the company’s AI systems autonomously carrying out such an act.

As per Google, the incidents arose from a case of mistaken identity during a “capture the flag” exercise on irregulars’ infrastructure. Gemini has been tasked with retrieving information from the software run by a fictional company inside the test environment, but that fictional company has the same name as a real company.

Although the model was not meant to have internet access, internet access was unintentionally unavailable. This enabled Gemini to access the real company’s systems, which in one case it did by guessing the passwords. But when the AI realised that it had accessed a real company, Google stated, it had stopped itself and exited.

Google has confirmed that it had been notified about the incidents in the month of July this year. But the company did not disclose at that time. The company thought these cases need not necessarily require public disclosure as the model ended the intrusion on its own.

ABOUT THE AUTHOR

Mary Sagaya

[email protected]