A Stanford study found that while leading AI chatbots scored above 90% on many real-time news questions, those results masked significant weaknesses in regional performance, source attribution and responses to flawed prompts.
Stanford Study Uncovers Flaws in AI Chatbots Answering News Questions