AI firms are watermarking generated text - heres why it wont work

AI companies have begun embedding watermarks in the output of their models to improve transparency and help crack down on misinformation, disinformation and cheating on homework, but there are limitations to this approach