India and U.S. in High-Stakes Talks Over Access to Anthropic’s Powerful “Mythos” AI

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 ...