Dino Dai Zovi: AI offensive powers demand non-software defenses, OAuth tokens won't cut it
dinodaizovi · x · 2026-10-12
Security expert Dino Dai Zovi responds to Matthew Green: the real challenge is that most orgs don't know how to prepare for exponentially improving offensive AI capabilities in such a short time.
Key points: threat-model against capabilities and environment knowledge once exclusive to top-tier nation states—though such attackers are intrinsically bound to software-based interfaces; identify reliable building blocks and build with non-purely-software defenses. Protecting modern Cloud/SaaS sprawl will be a huge challenge, and clonable OAuth bearer tokens everywhere won't cut it.
More from Safety
- Finding Vulnerabilities Is the Point: Debating AI Cyber Offense and Defense — teortaxesTex · 2026-10-24
- Security veteran asks: where would $1B and unlimited frontier-model tokens dent cybersecurity in 6-12 months? — chrisrohlf · 2026-10-24
- Musk says no H100s to rent; experts urge banning Chinese firms from renting US AI chips — chrisrohlf · 2026-10-24
- Samsung TV opens fake USB mic and records for exactly 5 seconds, test shows — ngxson · 2026-10-24
- Claude SSHes Into a Machine to Blink the Right Port When Asked Where to Plug In — peterjliu · 2026-10-24
- Security pros call AI security discourse incoherent: basics still win — AlexTensor · 2026-10-24