I recently wrote about cases of identity fraud in scientific publishing — including use of made-up personas as authors or peer reviewers and impersonation of real researchers. There are ongoing discussions on how to prevent such cases, but the solutions aren’t simple. www.nature.com/articles/d41...

Is AI sycophancy holding science back? A preprint by @jurafsky.bsky.social and his team found that AI chatbots are 50% more sycophantic than humans. In this piece, researchers tell me how AI sycophancy is creeping into many of the tasks that they use LLMs for. Read more @nature.com