AI Models Are Hallucinating Way More Than You Think
A comprehensive new AI benchmark reveals that nearly all major language models are more likely to hallucinate than provide correct answers. The findings challenge assumptions about AI reliability in real-world applications where factual accuracy matters most.