Tech: ChatGPT + 'scientific' data, Telsa flaws - known, but

Tech correspondent Mark Pesce joins Kathryn to look at a report in Nature that notes ChatGPT can be used to generate fake, but reasonably good-looking, scientific data to support a hypothesis. What does that mean for reviewers in how they approach data sets? And evidence presented at a court case suggests Tesla engineers - and the company's CEO Elon Musk - were aware of flaws with their auto-pilot system but failed to fix or update it. And a Spanish AI model is earning thousands from her "work" - and getting asked out by celebrities who don't realise she's not real.