Ministry of Electronics and Information Technology Secretary S. Krishnan said on Monday that India’s Computer Emergency Response Team (Cert-In) has operated an artificial intelligence war room with AI models that have 60-70 percent of the capabilities of frontier AI models such as Anthropic’s Claude Mythos and OpenAI’s ChatGPT 5.6.
“They (Cert-In) created a sandbox and a test environment where they used these AI models to try out different elements and identify possible vulnerabilities. So in a sense, we were able to do a dry run whenever we had access to these models,” Krishan said at the launch of Cert-In’s second edition of its Digital Threat Report.
According to Krishnan, the idea was also to prepare and check all vulnerabilities and potential patches when Indian companies and agencies have free access to these cutting-edge AI models.
Separately, he said the war room will ensure that a significant portion of the development of indigenous AI models and protection mechanisms takes place without relying heavily on generic models of these frontiers.
“Of course there are risks. The government is concerned that we should have access. Processing of sensitive elements will have to be done within government facilities,” Krishnan said.
In India, access to Anthropic’s latest general-purpose model, Claude Mythos, was initially extended to several government agencies, including the Indian Cybercrime Coordination Center (I4C), Cert-In, the National Critical Information Infrastructure Protection Center, and the Ministry of Telecommunications’ Digital Intelligence Platform as part of the expansion of Project Glasswing.
However, access was later suspended after the United States imposed restrictions on foreigners’ access to Claude Mythos. Krishnan said the Indian government is in talks with U.S. officials and companies to restore access to these models.
— Ashish Aryan, business standard
