India gets access to Anthropic's powerful AI cyber security model Mythos

Anthropic's Project Glasswing now covers organisations across more than 15 countries, including India, as AI firm broadens access to its advanced AI model for identifying cyber vulnerabilities

Artificial intelligence (AI) company Anthropic will provide access to its Mythos AI model to 150 organisations across more than 15 countries, significantly widening the availability of its advanced cyber security software beyond the United States and the United Kingdom.

 

Mythos is an AI model designed to identify cyber security vulnerabilities. Anthropic has previously described the technology as too powerful and potentially risky for unrestricted public release.

 

According to a report by the Financial Times, the company announced on Tuesday that it would expand “Project Glasswing”, an industry initiative that uses Mythos to detect and patch security vulnerabilities. The programme will now include companies and institutions across more than 15 countries.

 

 

Which countries have been granted access to Anthropic’s Mythos AI model?

 

India is among the countries newly granted access to the AI model, the Financial Times reported, citing sources.

 

The expansion also includes members of the Five Eyes intelligence alliance, namely Canada, Australia and New Zealand. Other countries joining the initiative include France, Germany, Italy, Switzerland, the Netherlands, Spain, Belgium, Sweden, Japan and South Korea, the Financial Times reported, citing people familiar with the matter.

 

The North Atlantic Treaty Organisation (Nato), headquartered in Brussels, has also been granted access to the technology.

 

Which organisations are part of Anthropic’s expanded Project Glasswing programme?

 

Among the companies receiving access are US identity and access management company Okta and South Korean technology groups Samsung Electronics, SK Hynix and SK Telecom.

 

The expansion also includes major financial market infrastructure providers such as Euroclear, Intercontinental Exchange and international payments network SWIFT. In addition, the European Union Agency for Cybersecurity (ENISA) is among the new participants, according to people familiar with the matter.

 

Anthropic said the latest batch of organisations includes operators of critical infrastructure spanning financial services, cyber security and technology. The company added that the expanded group also covers sectors that were underrepresented in the initial phase of the programme, including power, water, healthcare, communications and hardware.

 

Why is Anthropic widening access to Mythos?

 

Explaining its selection criteria, Anthropic said all participating organisations share a common characteristic: a successful cyber attack on their software systems could have severe consequences.

 

“What each partner has in common is that a successful attack on their codebase could be catastrophic,” the company said.

 

Anthropic estimated that, for most participating organisations, a major cyber attack could affect more than 100 million people, with significant implications for both national and global security.

 

The company also said it plans to broaden Project Glasswing further by prioritising additional essential infrastructure providers, maintainers of critical open-source software and safety-testing organisations worldwide.

 

The announcement comes a day after Anthropic filed for an initial public offering (IPO) that could value the company at more than $1 trillion.

 

Anthropic launched Claude Mythos Preview in April but initially restricted access to roughly 50 predominantly US-based organisations. The company cited the model’s advanced coding capabilities and the possibility that it could be misused for hacking.

 

The limited rollout had prompted concern among non-US organisations, including banks, regulators and governments, many of which sought access to Mythos or requested briefings on the vulnerabilities the company had identified.

 

Before the latest expansion, technology companies such as Microsoft, Apple and Oracle had already been granted access. The programme also included major US financial institution JPMorgan Chase and cyber security company CrowdStrike.