Anthropic is increasing Project Glasswing, its joint trade initiative to seek out and repair essential software program vulnerabilities utilizing AI, to about 150 new organizations throughout greater than 15 nations, the corporate said Tuesday.
The information comes a day after Anthropic mentioned it had filed confidentially for an preliminary public providing, following a $65 billion funding spherical at an almost $1 trillion valuation.
Anthropic’s Claude Mythos is on the coronary heart of Undertaking Glasswing. The AI agency dubbed the mannequin its strongest but, in a position to establish 1000’s of zero-day vulnerabilities over a number of weeks. In early April, Anthropic gave 50 initial partners, together with the U.S. authorities, entry to Claude Mythos Preview to scan their codebases for vulnerabilities and safety flaws.
The expanded listing of organizations with entry to Mythos as of at present covers energy, water, healthcare, communications, and {hardware} — industries that weren’t “well-represented” in Anthropic’s preliminary cohort, the corporate mentioned. Many who will now have entry are firms or nonprofits that keep codebases which different organizations and governments rely on, Anthropic famous within the weblog put up.
“What every associate has in frequent is {that a} profitable assault on their codebase could possibly be catastrophic,” the corporate mentioned. “For many companions, we estimate {that a} main assault may have an effect on greater than 100 million folks, with necessary ramifications for each world and nationwide safety.”
The expanded group contains organizations in nations pleasant to the U.S., together with Australia, Canada, France, Germany, Italy, Switzerland, the Netherlands, Spain, Belgium, Sweden, India, Japan, New Zealand, and South Korea, based on Financial Times, citing an individual acquainted with the matter.
The FT additionally reported a number of organizations which were given entry to Mythos, together with: U.S.-based id and safety administration device Okta; South Korean firms Samsung, SK Hynix, and SK Telecom; NATO, the U.S.-led army alliance headquartered in Brussels; and the EU’s cybersecurity company ENISA.
TechCrunch has reached out to Anthropic to verify.
Anthropic has mentioned it expects different AI firms to quickly develop fashions as succesful as Mythos Preview, which is why the agency is racing to determine safeguards inside Undertaking Glasswing.
Since releasing Mythos, rival OpenAI launched its personal cybersecurity-focused mannequin GPT-5.5-Cyber, which it has rolled out to a big group of companions for testing.
While you buy via hyperlinks in our articles, we may earn a small commission. This doesn’t have an effect on our editorial independence.