India in talks with US for 'equitable' access to Anthropic's Mythos AI model to secure critical infrastructure: Report

Indian government is actively negotiating with the US administration to secure “equitable” access to Anthropic’s recently unveiled Mythos AI model for Indian companies. According to a report by The Economic Times, the Centre is currently “looking for mechanisms” and working out “logistics” for Indian companies to get access to the model in order to ensure that the country’s critical infrastructure is not left vulnerable.

What is Mythos? Why does India want access to Anthropic’s latest AI?

Mythos is the latest AI model from Anthropic, which the company unveiled earlier this month but refused to release to the public owing to the cybersecurity risks it poses.

Also Read | Andy Mukherjee: Agentic AI could impact Indian real estate—here’s how

While talking about the capabilities of Mythos in its release announcement, Anthropic had said that its latest model is so capable of finding and exploiting software vulnerabilities that it could be used for activities that could have an adverse impact on economies, public safety, and national security.

Instead of releasing the model to the public, Anthropic had started a new initiative called Project Glasswing, under which it had given access to the model to around 40 organizations including Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorgan Chase, the Linux Foundation, Microsoft, and others.

Also Read | OpenAI ends Microsoft exclusivity, to offer AI models on rival cloud platforms

Notably, the list did not include any Indian company.

What is India doing to gain access to Mythos?

Finance Minister Nirmala Sitharaman, while recently speaking at an ET event, had confirmed that the Ministry of Electronics is in talks with the US administration, Anthropic, and vendors testing the new model.

“They are engaging with the US administration, Anthropic and with the vendors who have been given a chance to test (Mythos),” Sitharaman said at the event. “The cyber challenge we have because of the Mythos is going to be a big one….”

Notably, the government had also recently held meetings with banks and key cyber agencies last week in order to assess the risks posed by the Mythos AI model. Sitharaman had asked banks to strengthen their defenses and monitoring capabilities by engaging cybersecurity agencies and professionals.

The new ET report, while quoting unnamed officials, noted that the government doesn’t want to favour any particular company by granting access to some while excluding others. It also noted that conversations are ongoing with US authorities along with Anthropic officials.

Meanwhile, the report notes that the government asked the Indian Computer Emergency Response Team (CERT-In), the National Critical Information Infrastructure Protection Centre (NCIIPC), and the financial sector to move quickly to secure sensitive systems, including power grids, telecom networks, and banking infrastructure that could be vulnerable to the new model.

Also Read | Why did OpenAI ban its Codex AI from talking about ‘goblins’ and ‘gremlins’?

The government is also said to be considering a policy response as more companies begin releasing similarly advanced AI models.

“Currently, Anthropic has held off the wider release, but tomorrow more companies can launch such models,” the report cited an unnamed official as saying.

“They may release them without advance notice. The government needs to build its capacity as of yesterday,” it added.

A recent report by Axios had also noted that the US administration is exploring ways to allow US government agencies to work with Anthropic again after labelling the company a ‘supply chain risk’