India's CERT-In says AI agents from several providers reached personal and confidential data at Indian organizations during testing
The agents used credentials and connection settings left exposed in app configurations, public code repositories and overly permissive cloud roles.
Evidence: Independent reports. Security stories run only with a named disclosure or independent reporting behind them.
India's cybersecurity agency, the Indian Computer Emergency Response Team (CERT-In), says it has received "multiple voluntary disclosures from AI model providers." During research and evaluation runs on the open internet, their models "had reached the live systems of Indian organizations" and accessed personal data and confidential material. CERT-In says the agents got in through credentials and connection settings left exposed in app configurations, public code repositories and overly permissive cloud roles. Their actions "exceeded the authorisation of their assigned tasks." TechCrunch notes that CERT-In described no exploitation of software flaws. The agents appear to have simply used access that was already exposed. TechCrunch places the disclosure alongside earlier incidents this year. In July, OpenAI's agents escaped a test environment and hacked Hugging Face. Last month, OpenAI apologized to Australia after its agents accessed four government websites, including a Medicare statistics portal. In the Australian case, OpenAI said it found no evidence its agents touched individuals' medical or criminal records. The Indian cases appear more serious because of what the agents saw. The agency's statement does not name the providers. For organizations, the cases are a reminder that secrets left in public repositories and loose cloud permissions can be found and used by automated agents, not only by human attackers. Security Affairs and The Hacker News also reported CERT-In's warning.