I don’t see how this saves you anything time-wise. I would also be wary about letting AI change stuff and then being confident I would see all those changes while re-reading through. That is not how reading your own writing works, for me at least.
My eyeballs quickly refuse to eyeball. I proofread my thesis by using the inbuilt text to speech doodad in Microsoft word, because I am *that bad* at seeing what is in front of me and not what I think is there.
Yeah, I don't really think this is very different from word processor spelling and grammar checks. I think this is a reasonable permissible use for most circumstances, and I appreciate that they are up front about it.
That is the expected declaration statement for the use of AI. Far better to require disclosure than for this to be in the dark. Further, some scientists lacking English proficiency used to pay for language editing of their papers. AI can do this for free. Disclosure is a good thing!
There are also paid LLM tools that promise not to use your text for model training. Great tools for people for who English is only third or fourth language.
This is a good AI declaration. Says what they used and whether they accepted the suggested changes, and takes responsibility. If this was a post submitted to the academic blog I edit, we’d be happy with this.
Are the authors native English speakers? If not, then it's a fair use of AI, IMO. LLMs can be great tools for grammar correction, and can be of great help for those who do not speak English natively
Absolutely agree. Personally, I've not yet formed a firm view on many LLM/AI use cases, but I can see how this case is similar to a spelling/grammar checker. In fact, someone else replied with a similar comment.
Next it's going to be: "Declaration of AI Assistance in Citations and References: The authors declare we did not actually read the reference cited in this manuscript and instead relied AI generated summaries of these texts"
Comments