AI Chatbots Too Chatty on Government Queries, ODI Study Finds
Testing 11 large language models on 22,000 government service questions, the Open Data Institute found AI chatbots bury facts in waffle and make errors when told to be concise.
1 article tagged open-data-institute
Testing 11 large language models on 22,000 government service questions, the Open Data Institute found AI chatbots bury facts in waffle and make errors when told to be concise.