In addition to the considerations in this post , I add some further reflections on the risks of using ChatGPT and AI tools in general. Professor White, mentioned in the post, advises: 1) to use ChatGPT as a tool for “augmented intelligence” to increase, rather than reduce, our cognitive capabilities; 2) to question whether, because of the way we use AI, it will be easier in the future to replace us with AI itself or with someone who knows how to use it and, in addition, has competencies X, Y, and Z that we do not have or have lost. This is the real question that students who turn to ChatGPT or similar tools to ‘simplify their life’ should also ask themselves. The important issue is not whether by delegating a certain task to AI one might be ‘caught by teachers’, but whether in doing so, one is removing something from their own capabilities, increasing the likelihood, in the future, of becoming superfluous, at least from a work perspective. At the same time, however, educational institutions must ask themselves which old and new contents are significant in the era of AI. The ability to sum a long list of numbers, for example, is not particularly useful if calculators and PCs are available. What skills, then, is it important for students (and teachers) to develop when a myriad of AI tools are available? These will most likely be high-level skills, such as the ability to plan a task and formulate a prompt in a way that achieves a better result, to check the output, and to optimally integrate an AI system within a given context. But a high-level skill, which I believe is equally important, will also be to understand even better when not to use AI. We will talk about this in a future post.