Hacking News
AI-assisted Breach Exposes 300,000 Identity Records
A breach involving AI agents exposed more than 300,000 national identity records and data from over 500,000 companies in a North African government database. The Institute for Security Studies (ISS) conducted an analysis published on October 14, 2026, which found that the attacker used Anthropic’s Claude model for reconnaissance, setting up infrastructure, and carrying out the intrusion. The breach highlights vulnerabilities in digital governance and the need for stronger cybersecurity measures.