Security researcher: the best AI bugs live at the safety-security intersection
wunderwuzzi23 · x · 2026-10-08
Security researcher wunderwuzzi shares a takeaway from six years of AI security testing: the most interesting AI bugs he found typically sit right at the intersection of model safety and system security — where jailbreaks and conventional exploitation overlap. He also points readers to his long-running Machine Learning Attack Series, saying it's still worth a read.
More from Safety
- Utah becomes first US state to let AI prescribe medication without direct doctor review — NathanpmYoung · 2026-10-08
- Polymarket prices just 13% odds of a US AI safety bill by end of 2026 — Polymarket · 2026-10-08
- National Compute gifts $100M in compute credits to White House Genesis Mission — typewriters · 2026-10-08
- TheZvi: Automating Alignment Research Is Close to the Worst Possible Plan — TheZvi · 2026-10-08
- ChatGPT for Teens poses unacceptable risk as parental suicide alerts fail, new research finds — Polymarket · 2026-10-08
- Six years on, wunderwuzzi's ML attack series on model backdoors and Assume Breach still holds up — wunderwuzzi23 · 2026-10-08