Anthropic has announced it is expanding its Claude Mythos model to approximately 150 new organisations across the globe, bringing in new sectors including utilities and healthcare.
Claude Mythos is AI giant Anthropic’s cybersecurity focused model, which the company claims has found “thousands of high-severity vulnerabilities, including some in every major operating system and web browser”. Its rollout sparked fears in the global financial industry over their cybersecurity measures and whether they will be able to stay secure as AI becomes increasingly available to bad actors.
When it launched in April, the product was only available to a small number of companies that met Anthropic’s safety standards to ensure the technology was not open for abuse. Despite this, a small group of unauthorised users were able to access the programme in April.
The Tuesday announcement marks the first expansion of Project Glasswing, Anthropic’s industry initiative to use Mythos to find and fix security vulnerabilities. The company did not announce the organisations it was introducing individually, but said they cover “several industries that weren’t well represented in our initial cohort, such as power, water, healthcare, communications, and hardware”, many of which are vendors.
A person familiar with the matter told the Financial Times that the countries set to be granted access to Mythos include countries in the “Five Eyes” intelligence alliance, such as Canada, Australia and New Zealand. Other nations include France, Germany, Italy, Switzerland, the Netherlands, Spain, Belgium, Sweden, India, Japan and South Korea, the person said.
Anthropic intends to expand its geographical reach “much further” in the future, it added in the release.
