The Indian government is actively negotiating with the U.S. administration to secure “equitable” access to Anthropic’s latest AI model, Claude Mythos. This advanced model, described as a “watershed moment for cybersecurity,” has sparked global concern due to its unprecedented ability to identify and exploit software vulnerabilities.
The “Mythos” Threat and Opportunity
Anthropic has withheld Mythos from a general public release, citing risks to national security and global infrastructure.
-
Zero-Day Exploits: Testing shows Mythos can find and exploit “zero-day” flaws—vulnerabilities unknown to developers—across major operating systems and browsers at a rate 20 times higher than previous models.
-
Project Glasswing: Instead of a wide launch, Anthropic initiated “Project Glasswing,” granting early access to approximately 40 primarily U.S.-based organizations—including Google, Microsoft, and JPMorgan Chase—to help them patch defenses.
India’s Push for Inclusion
Currently, no Indian companies are included in the Glasswing program, leading to fears of a widening “AI divide”.
-
Economic Risk: Indian IT giants like TCS, Infosys, and Wipro worry that if U.S. competitors gain early access while they are excluded, it will create a severe competitive disadvantage.
-
National Security: Indian officials, including Finance Minister Nirmala Sitharaman, have flagged Mythos as a significant cybersecurity challenge for critical infrastructure like power grids and banking channels.
-
Government Action: The Ministry of Electronics and Information Technology (MeitY) is engaging directly with U.S. authorities and Anthropic executives to ensure Indian firms can build similar cyber resilience.
The Global Debate
While India seeks fair participation to protect its massive digital public infrastructure, the U.S. is reportedly exploring tighter controls to prevent the model’s capabilities from being misused or falling into the hands of adversaries. As these bilateral talks continue, Indian agencies like CERT-In and NCIIPC have been urged to accelerate their preparedness for the potential arrival of such advanced autonomous threats.
