If that’s the case, given the current publication process, the successive use of the tool by authors and then by reviewers could significantly slow things down by increasing the amount of work required from authors to satisfy the system. How do you see this issue?

But I’m wondering about a possible runaway effect that using such a tool might create. Suppose the paper is improved by taking into account these first comments, wouldn’t the AI still tend to identify major and minor gaps, even in revised versions?

Hi Oded, just tried it on a paper. I find the analysis overall very relevant, it’s impressive.

Thanks Matthias!

Thanks Ignacio

Thanks Lucas!