Mozilla used Anthropic's AI tool to find and fix 151 bugs in Firefox, signaling how developers will lean on AI for security work

Mozilla's Firefox team used Anthropic's Mythos (an AI system trained to spot security flaws in code) to scan Firefox's codebase and discovered 151 bugs that human reviewers had missed. The team then fixed these vulnerabilities before they could be exploited by attackers.

Unlike traditional security audits that require hiring outside firms to manually review thousands of lines of code, Mythos automatically scans for common patterns that lead to crashes or security holes — cutting the time and cost of finding these problems before they reach users.

Software developers across all industries will likely need to adopt similar AI-powered code review tools in the coming years, but the transition will be rough: teams must learn new workflows, understand when AI findings are accurate versus false alarms, and retrain security staff to work alongside AI instead of replacing them entirely.