It's not malicious in this case and readily apparent on the album.
I think it's fair to mention it in a review.
I agree it's not malicious, but it's still unprofessional. I wouldn't say it's readily apparent; I'd say it's possible, even probable, but to claim that the album is absolutely about Thom Yorke's divorce, even just a tiny part of it, is to make a massive assumption. A critic's job isn't to psychoanalyze the artist, it's to review the art as its own entity.
For another example, Braids' frontwoman Raphaelle Standell-Preston recently wrote
an essay about how the band's latest album was, at least in part, related to her own experiences with abuse. Now that she has actually unequivocally confirmed the relation between the album's content and her own experiences, it's fine to acknowledge that. The problem is that before she wrote the essay, people were jumping to conclusions, interviewers were grilling her about possible connections, especially in the song
"Miniskirt," and that's clearly not OK. Even if some people's assumptions regarding the song turned out to be accurate, it's still not acceptable to voice those assumptions in a professional environment without any sort of confirmation.
It's a different sort of scenario from Yorke's, but in terms of criticism and analysis it should be treated the same. I actually wrote a paper on "Miniskirt" in my senior year of college and I was very careful to make it perfectly clear that it was not appropriate to make any assumptions about the song potentially being autobiographical. If a college student can do that, surely a professional critic can too. And in fact, Pitchfork's review of Deep in the Iris did not jump to any such conclusions and I respect them for that. It's just a shame they didn't do the same here.