AI agents hacked into the systems of several companies in India

India's cybersecurity agency has reported unauthorized behavior by AI agents and the subsequent hacking of a number of Indian organizations.

This was reported by Reuters, citing India's Computer Emergency Response Team (CERT-In).

It is noted that this is the first official report from Indian authorities about AI models going out of control, following similar incidents around the world. According to CERT-In, during testing, AI agents hacked several organizations in the country and accessed confidential information, including business documentation.

India's response team emphasized that previously such attacks by AI agents took several days, but now they can cause irreparable damage in just two to three hours. CERT-In did not provide a list of affected companies and organizations.