Summary
AV-Comparatives evaluated 14 endpoint-security products against 50 multi-stage attack scenarios, including tooling built with AI-assisted development techniques. Eleven products achieved its highest Certified Leader status under a 92% pass threshold.
Why it matters
The test provides a concrete defensive benchmark for a threat model in which AI helps attackers build and operate intrusion tooling.
Related coverage
- Open Secure AI Alliance moves under Linux Foundation governance
- Gates Foundation commits at least $1 billion to equitable AI
- Elsevier and LG make chemistry in scientific images searchable