Eye On A.I. #313 Evan Reiser: How Abnormal AI Protects Humans with Behavioral AI
14 snips
Jan 16, 2026 
Evan Reiser, Founder and CEO of Abnormal Security, shares insights on the evolving cybersecurity landscape. He explains why phishing remains a primary threat, as attackers leverage human trust and sophisticated AI techniques. Evan highlights the limitations of traditional defenses, advocating for behavioral models that track 'known good' behavior. He emphasizes the AI arms race, the unique vulnerabilities posed by human judgement, and the intricacies of cybercrime economics, illustrating the urgent need for a deeper understanding of behavioral security.
AI Snips
Chapters
Transcript
Episode notes
Model Known Good Behavior
- Abnormal AI models "known good" behavior for every employee and flags deviations instead of relying on known-bad signatures.
- This approach catches unique, AI-crafted attacks that traditional signature-based defenses miss.
Humans Remain The Weakest Link
- Attackers exploit human trust and helpfulness because organizations must keep open communication channels.
- Tricking a support or accounts-payable employee is usually easier than breaching hardened infrastructure.
AI Lowers The Barrier To Sophistication
- Generative AI democratizes sophisticated social-engineering by automating personalized, high-quality phishing at scale.
- Criminals no longer need deep technical skills to craft targeted attacks using off-the-shelf models.