Indian government, tech firm join forces to test Anthropic’s Claude Mythos: Report

The recent unveiling of Anthropic's Claude Mythos, a model emphasizing cybersecurity, has stirred significant concern regarding its capabilities and potential for misuse across various sectors globally. In a strategic move, the Indian government has partnered with leading tech firms to rigorously test this model, aiming to uncover vulnerabilities in critical infrastructure. According to a report by Bloomberg, major Indian technology companies, including Infosys and Tata Consultancy Services (TCS), are poised to gain access to the Mythos Preview. These firms will be evaluating their software systems to detect possible threats and security weaknesses. Notably, Infosys plans to leverage Mythos to enhance security measures for its popular banking platform, Finacle. India's government cybersecurity agency, CERT-In, will also assess critical software systems, such as the Aadhaar national ID infrastructure and various government login platforms. However, it's important to note that many organizations do not have direct access to Mythos; instead, they are utilizing the Claude Opus 4.7 model to identify and rectify software vulnerabilities. Initially, access to Mythos was restricted to over 50 entities, including corporations and government bodies. Recent findings from Anthropic indicated that Mythos has successfully identified over 23,000 vulnerabilities, with more than 10,000 being classified as critical flaws during the early testing phases conducted by organizations like Cloudflare and Mozilla. Given India's pivotal role in the global banking and technology landscape, coupled with its rapid digital infrastructure expansion — highlighted by initiatives like Aadhaar, which aims to digitally register the country’s 1.4 billion citizens — the proactive identification and mitigation of cyber vulnerabilities is increasingly essential. In terms of testing logistics, the report details ongoing negotiations between India and U.S. officials to secure access to Mythos. The Indian government has earmarked a specialized testing facility and selected relevant sectors and experts to evaluate the model's performance and associated risks. U.S. Ambassador Sergio Gor stated, "India has requested access to various things, and as a trusted partner, we are definitely looking into those requests." As global governments advise financial institutions to bolster their cybersecurity frameworks and conduct regular vulnerability assessments, India’s Finance Minister Nirmala Sitharaman has emphasized the importance of vigilance against potential AI model threats, such as those posed by Mythos. With the ongoing tests, India aims to gain insights into its cybersecurity risks and capabilities, thereby enhancing its overall preparedness against cyber threats.