India gets access to Claude Mythos, Anthropic's powerful AI that found critical bugs in Linux, Firefox and OpenBSD

Anthropic is finally expanding access to its powerful Mythos AI model via project Glasswing. The company announced in a blogpost on Tuesday that it will be bringing Mythos to 150 additional organizations across 15 countries, including India.

“For most partners, we estimate that a major attack could affect more than 100 million people, with important ramifications for both global and national security.” Anthropic warned in its latest blog

Anthropic said that with AI-powered cyber capabilities advancing rapidly, a Mythos-class model could be released by other AI companies in the next 6 to 12 months. The Claude maker says these models would be released ‘without safeguards that prevent misuse.’ which would lead to cyberattacks occuring much more often and in much more unpredictable forms.

Anthropic brings Mytos to India:

While Anthropic did not officially mention 15 countries it is expanding Project Glasswing to, a report by Financial Times that the company is expanding Mythos to countries in the “Five Eyes” intelligence alliance. Apart from India, this includes Canada, Australia and New Zealand. Other nations include France, Germany, Italy, Switzerland, the Netherlands, Spain, Belgium, Sweden, Japan and South Korea

In its blogpost, Anthropic says that the companies covered in its latest batch cover several industries such that weren't earlier well represented in the Mythos rollout including power, water, healthcare, communications, and hardware

“And many of the new partners are vendors—companies or nonprofits that maintain codebases that are relied upon by lots of other organizations around the world, including governments.” the company wrote

However, no specific company name was mentioned by Anthropic. The FT report does note that companies receiving access to Claude Mythos Preview include identity management firm Okta, South Korean tech giants Samsung, SK Hynix and SK Telecom, financial infrastructure providers such as Euroclear, Intercontinental Exchange (ICE), which owns the New York Stock Exchange, and international payments network Swift.

The report further states that Nato, the US-led military alliance, and Enisa, the European Union's cybersecurity agency, have also been granted access to the model.

What is Project Glasswing?

Notably, the San Francisco based AI company had launched Project Glasswing in April when it announced Mythos and refused to release the model to public because of the cybersecurity risks that the model posseses. Instead, it decided to give access to around 50 organizations like Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, and NVIDIA to find and fix bugs present in their technology.

The company had also revealed that Mythos was able to uncover thousands of previously unknown software vulnerabilities that had gone unoticed for decades in the some of world's most widely used systems.

For instance, Anthropic had said that MMythos identified a 27-year-old vulnerability in OpenBSD, an operating system widely used in firewalls and other critical infrastructure. The powerful AI model also discovered 271 vulnerabilities in Firefox.