Researchers at Carnegie Mellon have figured out how to make AI models like ChatGPT serve up prohibited material by sending it nonsensical text strings…sort of like a buffer overflow or SQL injection attack.
Advertise here with Carbon Ads
This site is made possible by member support. ❤️
Big thanks to Arcustech for hosting the site and offering amazing tech support.
When you buy through links on kottke.org, I may earn an affiliate commission. Thanks for supporting the site!
kottke.org. home of fine hypertext products since 1998.
Stay Connected