The Indian Computer Emergency Response Team (Cert-In) has operationalised an artificial intelligence war room to test AI models that carry 60-70 per cent of the capability found in frontier systems such as Anthropic’s Claude Mythos and OpenAI’s ChatGPT 5.6, according to S Krishnan, secretary at the Ministry of Electronics and Information Technology. Krishnan disclosed the move on July 13, 2026, at the launch of the second edition of Cert-In’s Digital Threat Report.
According to Krishnan, Cert-In has built sandboxes and testing environments to experiment with these AI models and surface possible vulnerabilities, effectively running what he called a dry run for the day India gains fuller access to frontier AI systems.
The exercise is also meant to identify potential patches and safeguards that Indian companies and government agencies would need once unrestricted access to frontier models becomes available, Krishnan said.
Beyond vulnerability testing, the war room is intended to support the development of indigenous AI models and protection mechanisms, reducing India’s reliance on general-purpose frontier systems built abroad, he added.
Krishnan acknowledged the risks involved, saying the government wants access to these models but insists that processing of sensitive data take place on premises within government infrastructure rather than externally.
Access to Claude Mythos, Anthropic’s latest general-purpose model, had earlier been extended to select government agencies, including the Indian Cyber Crime Coordination Centre, Cert-In, the National Critical Information Infrastructure Protection Centre and the Department of Telecommunications’ Digital Intelligence Platform, as part of the expanded Project Glasswing. That access was subsequently halted after the United States imposed restrictions on foreign access to Claude Mythos, though Krishnan said the government is in discussions with US counterparts and the companies involved to restore it.