Govt plans empowered and unified command centre to tackle AI risks

New command centre would be tasked with ensuring timely implementation of the latest cybersecurity measures

The emergence of highly powerful artificial intelligence (AI) models, which are proving to be a double-edged sword, has triggered action-oriented discussions within the government. It is learnt that the Centre is likely to set up an empowered and unified centre of command to address cybersecurity issues arising from these advancements in AI and general-purpose AI models such as Anthropic’s Claude Mythos, which can perform several tasks without needing separate training for each. 

This new command centre, expected to function under the Ministry of Electronics and Information Technology (Meity), would be responsible for driving the government’s overall policy response to emerging AI threats and coordinating with various departments across central and state governments for their preparedness. 

It would also be tasked with ensuring that the latest cybersecurity measures are put in place well within stipulated time periods, two officials with knowledge of the development said. 

The new body will likely have representatives from the government, academia and frontier AI firms, and would be empowered by a new AI law currently being prepared by the IT ministry, a senior government official said, requesting anonymity. Frontier AI refers to the most advanced, cutting-edge general-purpose models. 

The need for such a unified cybersecurity command centre was felt because the government believes the reaction time to threats arising from general-purpose models and frontier AI will shrink drastically over the next few months, the official said.

 

“The government has to act as one unit since the threat from these newer models will soon scale up,” this person added. 

The government has also reached out to the handful of companies that still have access to Claude Security, and is encouraging them to share the results of using it with the industry, the second official said. 

“The tool (Claude Security) is in the public Beta phase right now, so the access is tightly controlled,” the official said. “All the participating organisations are vetted, they must meet strict security requirements, and even then, access is limited to designated users within the company. But we have been talking to the companies to understand what data they can share,” the official added. 

An email to MeitY did not elicit a response. 

The development comes weeks after Indian organisations, including the Indian Cyber Crime Coordination Centre, the Indian Computer Emergency Response Team, the National Critical Information Infrastructure Protection Centre, the Digital Intelligence Platform of the Department of Telecommunications, besides select teams of some Indian IT services companies, were granted access to Anthropic’s Claude Mythos and other latest AI models from the company. 

The access was, however, withdrawn after the United States government prohibited Anthropic from providing access to Fable 5 and Mythos 5 to any foreign national, whether inside or outside the US, including the company’s foreign employees. 

In a blog published on June 3, Anthropic pointed out that malicious actors had been using AI in ways that made them more dangerous. “More specifically, threat actors are using AI in the later, more complex stages of their cyber operations. Cyberattacks are becoming more autonomous, and the fact that AI can be used to chain together many parts of the attack means that the old ways of differentiating high- from low-risk actors are no longer as effective,” the blog read.