Mike explains a paper from the University of Maryland, proposing a neat trick to 'watermark' the output of large language models such as ChatGPT. Dr Mike Pound is an image analyst at the University of Nottingham.
