@JASPStats
When considering whether or not the use of commercially available LLMs is ethical, perhaps it would be good to take a broader perspective. The first that comes to mind is energy consumption: are the claimed improvements on this text worth the use of such a power-hungry technology? This might not be visible for single cases but, at scale, the negative impact is well documented (e.g., for the environment as well as for communities who have to deal with newly built data centers). If one takes this into consideration (and there are more issues), is it ethical to claim that not using commercially available LLMs is unethical?
@antonioschettino I agree that energy consumption should not be frivolous. I don't think its use for improving academic writing is frivolous (at all). The extent to which use of LLMs are more taxing than, say, streaming a movie, using Google, or driving a car is unclear to me (internet gives me conflicting info; maybe I should ask an LLM ;-))

@EJWagenmakers the internet (and LLMs, since they are largely trained on the same data) gives you conflicting information because tech companies are far from transparent on this topic. Whether or not the energy consumption is higher or equivalent to the other things you mentioned (which, we already know, require energy mostly from fossil fuels) might overlook the fact that people didn't stop all these other activities, they added the use of LLMs on top of that.

So, to me, the question remains: is the use of LLMs ethical, in the sense that the advantages outweigh the costs? At the moment, my answer is no. This also applies to scientific writing: while it might be a less frivolous use of LLMs than asking to generate a picture of a rat on a unicycle, it still does not justify the cost.

@antonioschettino To answer your question we would first need to know the costs. I was busy correcting my writing instead of watching a movie on Netflix...