Mozilla has moved AI-assisted bug hunting from the margins to the center of its security playbook.

The Firefox developer says it has “completely bought in” on AI-assisted bug discovery after Mythos identified 271 vulnerabilities, according to reports. Just as notable, Mozilla says the findings came with “almost no false positives,” a claim that cuts to the biggest concern around automated security tools: whether they waste engineers’ time with noisy, unreliable alerts.

Key Facts

  • Mozilla says Mythos found 271 vulnerabilities.
  • The company reports “almost no false positives” in those findings.
  • Mozilla says it has “completely bought in” on AI-assisted bug discovery.
  • The effort centers on software security work tied to Firefox development.

That endorsement matters because bug discovery lives or dies on trust. Security teams already juggle huge volumes of data, and tools that flood them with weak leads often lose support fast. Mozilla’s comments suggest Mythos did more than surface theoretical issues; it appears to have produced findings engineers considered actionable enough to change the company’s posture on AI in vulnerability research.

Mozilla says AI-assisted bug discovery found 271 vulnerabilities with almost no false positives, a result strong enough for the company to say it has “completely bought in.”

The shift also reflects a broader change across the software industry. Developers have spent years testing AI for code generation and productivity, but security presents a tougher standard. A useful tool must spot real weaknesses before attackers do, and it must do so without drowning teams in false alarms. Mozilla’s embrace of Mythos suggests at least some AI systems now clear that bar in practical engineering work, not just demos or research papers.

What comes next matters beyond Firefox. If Mozilla expands this approach, other software makers will face more pressure to adopt similar systems for code review and vulnerability detection. That could reshape how companies secure products, how fast they patch flaws, and how they measure trust in AI tools. For users, the immediate takeaway is simple: one major browser maker now sees AI not as a side experiment, but as a serious part of defending widely used software.