AI and the Limits of Scientific Precision

The use of AI language models such as ChatGPT to summarise scientific studies carries a considerable risk: the results of the studies analysed are often overinterpreted or outright generalized. This is the central result of a recent study conducted by Uwe Peters (University of Utrecht) and Benjamin Chin-Yee (University of Cambridge and Western University), who have been intensively concerned with the accuracy of such AI summaries (via press rele…