AI chatbots oversimplify scientific studies and gloss over critical details -- the newest models are especially guilty

More advanced AI chatbots are more likely to oversimplify complex scientific findings based on the way they interpret the data they are trained on, a new study suggests.