A study found people who sought information via a generative AI chatbot could see lower learning outcomes than those using a traditional search engine. Researchers at institutions including Georgia Tech and the University of Michigan compared ChatGPT with Google Search in an eight-day experiment on nutrition and diet planning. The ChatGPT group tended to perform worse, especially in applying knowledge and comparing and evaluating information, and also reported greater dissatisfaction during learning.

A Reuters review found Tesla’s promoted safety statistics for its Full Self-Driving (FSD) system were based on a comparison using different crash measures. Researchers and former Tesla staff interviewed were sceptical of claims that FSD is up to 10 times safer than human drivers. Some former data labelers and an autonomy engineer said they would not trust the system. The review also cited reports of testing and mapping practices and growing regulatory scrutiny.