Mozilla Deploys Anthropic AI to Patch Hundreds of Firefox Vulnerabilities

4 articles · Updated · Engadget · Apr 21

Mozilla has used Anthropic’s Claude Mythos AI model to identify and patch 271 security vulnerabilities in the latest Firefox 150 release.
The AI model reportedly found far more bugs than previous automated tools, helping Mozilla address issues that would have required extensive human effort.
This marks a significant shift in cybersecurity, as advanced AI tools may help defenders close the gap with attackers in identifying software vulnerabilities.

Will AI's power to find flaws crush open-source projects under a wave of unfixable bugs?

Now that AI can find decades-old flaws, is any software on Earth truly safe from exploitation?

As AI hardens our code, will hackers simply pivot to exploiting human error more effectively?

Is the AI security revolution real, or are social engineering attacks still the biggest threat?

Who will pay for the massive global code cleanup that AI has just made necessary?