Govt plans 'AI-against-AI' approach to strengthen digital public infra
Jun 1, 2026
Govt is considering deploying artificial intelligence to identify vulnerabilities and strengthen digital public infrastructure as concerns grow over advanced AI models' cybersecurity capabilities
The government has planned to deploy artificial intelligence (AI) to strengthen India’s digital public infrastructure and address gaps in the cybersecurity system, amid growing concerns that advanced AI models could make it easier for cyber criminals to exploit critical systems, according to two finance ministry officials aware of the matter.
“We have planned an AI-against-AI approach,” one of the officials told Business Standard.
The move comes at a time when Anthropic has claimed that its latest advanced generative AI model, Claude Mythos, has the potential to locate dormant vulnerabilities hidden in decades-old code and exploit them.
“Manual audits are not enough to fill all gaps; we need to deploy AI to scan the system for older dormant bugs,” the second official said.
Claude Mythos is being viewed globally as a potentially disruptive cybersecurity development because of its reported capabilities. According to Anthropic, the model can identify and exploit zero-day vulnerabilities across major operating systems and web browsers when directed by a user to do so.
The Ministry of Finance, Reserve Bank of India (RBI), and the Ministry of Electronics and Information Technology (MeitY) are working as a task force to strengthen India’s cyber defence system, the official said. “RBI and the finance ministry are also working to ensure banks take pre-emptive measures to secure their information technology systems and safeguard financial resources,” one of the officials said.
“MeitY is in touch with Anthropic and the US government for early access to the AI model,” the official said. “We also need to verify whether the Mythos AI model is actually capable of what Anthropic claims,” the official added.
According to the officials, the bigger worry is that powerful AI models could dramatically lower the skill required for cyber criminals and hostile actors to discover and exploit vulnerabilities across critical financial and digital infrastructure.
Both officials spoke on the condition of anonymity as the discussions are private. The finance ministry did not respond to a request for comment.
At the same time, one of the officials stressed that India’s digital public infrastructure is already strong and layered. “Just a year ago, our IT infrastructure withstood countless cybersecurity attacks during Operation Sindoor,” the official said.
India’s digital public infrastructure handles billions of transactions and underpins welfare transfers, retail payments and banking services. This makes cybersecurity vulnerabilities a potential financial stability concern for India rather than merely an IT issue.
Finance Minister Nirmala Sitharaman had convened a high-level meeting with heads of banks in April to assess the likely cybersecurity risks related to AI models. Meanwhile, public sector banks have also decided to step up spending on IT infrastructure and cyber resilience.
Currently, Anthropic has extended access to Mythos Preview to select US companies, including Amazon Web Services, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, The Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks under its initiative called Project Glasswing.
As part of Project Glasswing, the launch partners are using Mythos Preview as part of their defensive security work. Anthropic has also granted access to a group of more than 40 additional organisations that build or maintain critical software infrastructure so they can use the model to scan and secure both first-party and open-source systems.
Following its pilot launch in April, Anthropic said Mythos Preview had already uncovered thousands of high-severity vulnerabilities, including some affecting every major operating system and web browser.
[The Business Standard]
