The Canadian Centre for Cyber Security is among the organizations newly authorized to use the advanced AI model, AI Minister Evan Solomon said, as the U.S. AI company widens the circle of companies and officials allowed to work with the model that’s startlingly good at exploiting software vulnerabilities. He called it a “very important step to increase our access and to make sure that Canadian systems are protected.” (The Logic)

Talking point: The model has found bugs that have lurked unnoticed in commonly used software for many years and can chain them together into effective hacking tools. Anthropic publicly announced Mythos’s existence in April, along with what it called Project Glasswing—a joint effort with major software companies that lets them use the tool to find and fix exploitable bugs in their own products. On Tuesday, the company expanded the list of Glasswing participants from about 50 to about 200. Anthropic has not responded to The Logic’s inquiries about which Canadian entities are involved and Solomon said he wouldn’t reveal any.