A note on my use of AI

I use generative AI in my academic work, mainly when revising my own texts. Once I know what I want to say, it helps me say it more fluently: it suggests wording, smooths out sentences and lets me spend less time on language and more on content. Occasionally, I also use it to test whether an argument holds up.

I see no reason to do without this help just because I could manage without it. I don’t think scientists should be judged by how elegantly they write. A text that reads well is easier to follow, but what makes research good is the work behind it: the questions, the reading, the data, and whether the conclusions are sound.

This is also where AI reaches its limits. It can make a sentence smoother, but it cannot decide what the sentence should claim, and it is not always precise where precision matters most. So the judgement remains mine. I read the literature I cite myself, I am able to explain and defend what I publish, and I take responsibility for what appears under my name.

There is one part of my work where I do not use AI at all: peer review. Manuscripts under review are confidential, and they stay out of AI tools.

Within these limits, I consider AI a useful tool for my writing.