Jason Koebler18h ago

New: Researchers have quantified how easy AI search is to manipulate. Just 13 words buried in a random Reddit comment can poison AI search results. They suggest this is not easy to stop: "The way you can attack these systems is so much dumber than you think it is"

https://www.404media.co/it-is-trivially-easy-to-use-reddit-to-manipulate-ai-search-research-suggests/

It Is Trivially Easy to Use Reddit to Manipulate AI Search, Research Suggests

"We show that a tiny snippet—just 13 words—of retrieved text on a UGC website like Reddit, Wikipedia, Quora, or Facebook can change AI agents to output spam / scam content pretty consistently."

404 Media
Show threadtuv
@jasonkoebler On the one hand, LLMs need vast amounts of data, they won't learn a thing if they only get all the information a human receives in a life time; on the other hand, they are unable to ignore the tiniest misdirection they receive
Jun 15 at 3:37pmWeb