AI & Tech
OpenAI's new AI found 81 serious Firefox bugs
A cybersecurity model for approved researchers is finding and helping fix software holes at a scale humans could not.
In short
81previously unknown high-severity Firefox bugs found
- 81 Firefox bugs found
- 948 Linux kernel flaws
- Access limited to approved researchers
What happened
OpenAI launched GPT-6 Cyber, a model only for authorised security researchers. Using it, researchers reported 81 previously unknown high-severity bugs in the Firefox browser and working exploits for 948 Linux kernel vulnerabilities — with fixes released or in progress.
Access is limited at first through a programme called Daybreak Red.
Why it matters
The same tools that help defenders find holes could help attackers too, which is why access is restricted. For everyday users, the takeaway is simple: keep your browser and phone updated, because many of these fixes arrive as routine updates.
