Mozilla announced on Tuesday that its Firefox 150 browser release includes protections for 271 vulnerabilities identified using early access to Anthropic’s Mythos Preview AI model, marking a significant test case for how artificial intelligence could transform cybersecurity in 2025.
The Firefox team gained access to Mythos Preview through direct collaboration with Anthropic, though Mozilla is not formally part of the company’s larger consortium, Project Glasswing. Both Anthropic and OpenAI have announced new AI models in recent weeks with advanced cybersecurity capabilities, releasing them only to limited private partners while convening industry working groups to assess their impact.
Bobby Holley, Firefox’s chief technology officer, said the AI tools represent a dramatic shift in vulnerability hunting. For years, organizations relied on a combination of automated techniques like software fuzzing and manual research to find flaws. Now, he says, automated AI techniques can cover “the full space of vulnerability-inducing bugs” that previously required expensive human analysis.
The shift creates what Holley calls a “bootcamp” that all software must undergo to find and fix latent vulnerabilities before these AI capabilities become widely available to attackers. “Every piece of software is going to have to make this transition, because every piece of software has a lot of bugs buried underneath the surface that are now discoverable,” he said.
Holley acknowledged the transition requires significant resources and discipline to manage the “firehose of bugs” AI tools can uncover, but said this effort is necessary given that attackers will inevitably gain access to similar capabilities soon.
The development could particularly impact open source software, which is often maintained by small volunteer teams or single individuals, and “abandonware” that is no longer maintained at all.
Source: Business Latest