Generative AI succumbs to conversational misinformed pressure and argument. ChatGPT 3.5 was most vulnerable to reaffirming misinformation during a conversation containing repeated false statements; Claude 3.5 Sonnet was the least. All seven were more susceptible to misinformation on obscure topics.
Read Full Article
#GenerativeAI #Misinformation #ChatGPT3_5
Read Full Article
#GenerativeAI #Misinformation #ChatGPT3_5
The University of Arizona
Study: Generative AI succumbs to conversational misinformed pressure and argument
U of A research assessed seven different generative AI language learning models. Their work reveals intrinsic limitations that might go undetected during one-off interactions.