Centre In Talks With US, Anthropic For Access To Claude Mythos: Report

Why It Matters

Access to Claude Mythos is crucial for safeguarding India’s critical digital infrastructure against sophisticated cyber threats, and the negotiations set a precedent for AI governance and security collaboration.

Key Takeaways

• India negotiating with US and Anthropic for Claude Mythos access
• Model can uncover decades‑old software vulnerabilities across critical sectors
• Finance ministry urges banks to report threats and boost cyber defenses
• Nasscom seeks inclusion of Indian firms under Project Glasswing
• Anthropic currently limits model to non‑Indian companies, citing security risks

Pulse Analysis

Anthropic’s Claude Mythos, launched in April, is a generative AI system tuned for advanced coding and cybersecurity tasks. Unlike typical language models, Mythos can automatically scan codebases, identify zero‑day flaws and even craft exploit scripts, exposing vulnerabilities that have lingered for years in operating systems, browsers and industrial control software. This capability places the model at the intersection of innovation and national‑security risk, prompting regulators worldwide to treat it as a dual‑use technology that could be weaponized by hostile actors.

The Indian government has moved swiftly to mitigate those threats. Finance Minister Nirmala Sitharaman flagged Mythos as a “big cyber challenge” and instructed the Ministry of Electronics and Information Technology to coordinate with the United States and Anthropic on a secure access framework. Parallel meetings with the RBI, major banks, the Indian Banks’ Association and cyber‑security agencies resulted in directives for immediate incident reporting to CERT‑In and accelerated hardening of power‑grid, telecom and banking networks. Nasscom’s Project Glasswing petition seeks to embed Indian firms in the model’s rollout, arguing that domestic codebases are integral to global supply chains.

Globally, the Mythos debate underscores a shift toward tighter AI export controls and collaborative governance. Countries that grant access are expected to impose stringent usage audits, data‑localisation clauses and real‑time monitoring to prevent malicious exploitation. For Indian technology companies, securing a vetted channel could unlock powerful development tools while obligating them to adhere to heightened compliance regimes. The outcome of these negotiations will likely set a benchmark for how emerging AI models are integrated into critical‑infrastructure protection strategies across emerging markets.