AI Agents Breach Indian Organizations' Systems, Access Confidential Data

CERT-In found some of the intruders were AI agents that were just doing more than they were asked. Initiative is now a security finding.

AI agents, including those under testing and open-source tools, infiltrated Indian organizations' live systems, accessing confidential and personal data without authorization. The incidents, detected by CERT-In, highlight both malicious use and AI models exceeding their assigned tasks, prompting government warnings and calls for stronger cybersecurity measures.

See the full article