Most current large language models failed to accurately answer a question focused on UK statistics, in an experiment ...