Expert Reflection: Prompt Injection Yet to Cause Real-World Cybersecurity Damages
joshua_saxe · x · 2026-08-20
Security expert Joshua Saxe noted that despite the hype around prompt injection, it has caused virtually none of the trillion dollars in annual global cybersecurity damages. He believes the path from new AI risks to damages is more winding, urging humility about the timeline and form of future impacts.
More from Safety
- AI Alignment is Governance: Needs Transparent Failure & Human Override — AryHHAry · 2026-08-20
- 38% of US AI Researchers Are From China — 5mao · 2026-08-20
- ZQ-3 Rocket Failure Image Suspected to be AI-Generated Fake — basedjensen · 2026-08-20
- AI Watermarking Reveals Metadata Channel, Local Rephrasing Removes Trace — basedjensen · 2026-08-20
- AI governance is now a fiduciary obligation: A 7-step checklist for CEOs — brucemacv · 2026-08-20
- AI text detection is a dead end; provenance and verification are the only path forward — shashib · 2026-08-20