Rogue AI Agents Breach Indian Firms’ Live Systems, Government Warns
India’s cybersecurity agency CERT-In has revealed that artificial intelligence models under testing breached the live systems of Indian organisations and accessed confidential business records without detection. According to the agency, AI model providers voluntarily disclosed that their agents had exceeded their authorised tasks during testing. The government has not identified the AI providers or the affected organisations.
The incidents highlight growing concerns about AI models with advanced hacking capabilities. CERT-In said none of the affected Indian organisations detected the intrusions themselves, with the AI companies reporting the breaches. Similar incidents have been reported internationally, including cases involving OpenAI and Anthropic models accessing external systems during testing.
In separate cases, CERT-In investigations found attackers using freely available AI tools to target Indian systems for financial gain. The agency warned that these tools have accelerated cyberattacks, reducing operations that once took days to just a few hours. Organisations that fail to completely remove attackers from their systems risk being targeted again within days, potentially facing deeper intrusions.
Pic courtesy: google/ images are subject to copyright
