Eugene Alvin Villar on Nostr: Wow! This is a pretty dumb implementation of LLMs. So in the course of doing a fact ...

Wow! This is a pretty dumb implementation of LLMs.

So in the course of doing a fact check, the fact checkers will state the mis/disinformation. The AI models then think that the authoritative fact checkers actually claim the wrong fact.

https://fullfact.org/online/amazon-alexa-misleading-voice-assistant-more-answers/

#AI #GenerativeAI #LLMs #AIFails