From that ArsTechnica article: "Every statement of fact" in ChatGPT's SAF case summary "pertaining to Walters is false," Walters' complaint said. Well, sure. It's well-known that ChatGPT just makes stuff up. Don't they realise that it can't access the Internet, and it can't even see the source text of its training data? Maybe people should learn a bit about how GPT works before they attempt to use it for anything more than an entertaining toy...