A team of artificial intelligence specialists at Google’s DeepMind has developed an AI-based system called SAFE that can be used to fact check the results of LLMs such as ChatGPT. The group has published a paper describing the new AI system and how well it performed on the arXiv preprint server.