There has been a lot on social media recently about the remarkable AI 'writer' ChatGPT . Many of my contacts are either academics or writers - and in both fields there has been concern about the ability of such AI software to interfere with their work. As far as academics are concerned, the software arguably calls into doubt the value of the essay as means of assessing student knowledge (particularly the short exam-style essay, which some universities still allow to be submitted online). And some writers have worried that their jobs are at risk if, for example, a news outlet can get ChatGPT to knock up a story for free. My assessment is that the software is extremely impressive - but it has limitations that are likely to continue for a long time. Even with these limitations, it certainly could make inroads into student essays and poorly churned out journalism, but is unlikely to be a significant threat if some degree of insight or fact checking is required. Let's take a lo