India Gets Access to Anthropic’s Mythos AI Model Under Project Glasswing

IPO-bound AI giant Anthropic has expanded access to its cybersecurity AI model Mythos under ‘Project Glasswing’ to around 150 organisations across more than 15 countries, including India.

The initiative pairs the closed Claude Mythos Preview model with participating organisations to proactively identify and fix critical zero-day vulnerabilities before attackers can exploit them. 

The model, designed to detect software security flaws, was initially subject to tight restrictions after early testing showed it could rapidly surface thousands of vulnerabilities, raising concerns about potential misuse after its April rollout.

At first, Project Glasswing was limited to about 50 partner organisations, including major tech firms such as Amazon, Google, Microsoft, Apple, and NVIDIA, cybersecurity companies like CrowdStrike and Palo Alto Networks, and early access participation from the UK’s AI Security Institute.

By late May, these early users leveraged Mythos to identify over 10,000 serious security vulnerabilities that could be exploited by attackers, Anthropic said.

Besides India, organisations from Australia, Canada, France, Germany, Italy, Switzerland, the Netherlands, Spain, Belgium, Sweden, Japan, New Zealand, and South Korea will have access to Mythos as part of the latest expansion, Financial Times reported.

The move comes as Anthropic positions Mythos as a cybersecurity tool for governments, critical infrastructure operators, and key institutions, even as concerns grow globally over its ability to identify software vulnerabilities and the risks it may pose.

India Assesses AI Cybersecurity Risks 

Amid these concerns, India also stepped up efforts to identify software vulnerabilities. Days ago, reports said that India was testing key financial and government software to assess vulnerabilities against Mythos AI model. Major tech firms like Infosys and Tata Consultancy Services were said to be conducting controlled security tests, while CERT-In was reviewing critical systems such as Aadhaar and government login platforms.

As these organisations didn’t have access to Mythos, they were using Anthropic’s Claude Opus 4.7 to identify and fix vulnerabilities.

Meanwhile, the central government is reportedly pushing for sovereign hosting of the Claude AI model in India, citing jurisdictional, compliance and national security concerns around foreign-hosted infrastructure used in sensitive sectors such as banking, telecom and critical infrastructure.

Anthropic also held meetings with officials from the finance ministry, MeitY and CERT-In regarding access to the Mythos model.

In May, SEBI formed a task force called cyber-suraksha.ai to assess cybersecurity risks from Mythos. The group will develop a unified mitigation framework, share threat intelligence, improve vulnerability management, strengthen response playbooks, monitor incidents, review third-party vendors and assess the security posture of regulated entities.

SEBI said AI-based vulnerability tools like Mythos could increase risk exposure and raise concerns around data confidentiality and output reliability. Following consultations with market participants, the regulator also issued cybersecurity guidelines for regulated entities, including immediate patching, AI-based vulnerability assessments, stronger vendor coordination, stricter API controls, enhanced monitoring, zero-trust frameworks and onboarding onto market security platforms, along with long-term AI-based threat detection plans.

The Centre was also in talks with the US and Anthropic to explore access mechanisms for Indian companies while safeguarding critical infrastructure. 

Earlier, finance minister Nirmala Sitharaman also flagged the risks posed by the model. Recent meetings involving Sitharaman, IT minister Ashwini Vaishnaw, banks, the RBI, MeitY, the NPCI and other agencies focused on strengthening cybersecurity defenses and coordination. Banks were directed to report cyber incidents quickly and work with CERT-In and other agencies.

The RBI is also reviewing the risks posed by Mythos in coordination with global regulators and domestic stakeholders. 

The post India Gets Access to Anthropic’s Mythos AI Model Under Project Glasswing appeared first on Inc42 Media.