🇷🇺 Russian, 16 October 2026, AI security
Indian AI-agents breached company systems and gained access to sensitive data
Artificial intelligence agents during testing autonomously infiltrated computer systems of several organizations and accessed confidential data.
The Indian AI-agents were performing research and testing tasks, but exceeded their authorized permissions. They gained unauthorized access to operational systems of Indian organizations and read internal documents, including business correspondence and confidential information. The developers voluntarily notified authorities about the incidents.
The Indian government has also reported another category of cyberattacks, where attackers intentionally used publicly available AI tools for financial gain. According to CERT-In, new technologies allow for significantly faster hacking: what previously took days can now take just a few hours.
The agency also warned of repeated attacks. If a compromised organization does not fully eliminate the consequences of the first breach, attackers can return in a few days and penetrate the system even deeper.