Scott Alexander is holding a book review contest, and the usual procedure is that ~10 reviews are selected and posted on the blog. This question resolves YES if any of those are confirmed to have been written primarily by an AI. If there's no such evidence by the end of the year, this market will resolve NO.

If a review is posted with a disclaimer "this is written by ChatGPT", that'll count. It doesn't need to be a surprise revelation two months later.

The review doesn't need to have been written exclusively by the AI. This question is about whether the AI is the primary author. (If the review was submitted to a journal where author order matters, who would be listed first?) The word of the submitting human will ordinarily be sufficient for this. A plausible workflow that would make this resolve YES:

• Human reads book, has list of 6 important points to make.

• Human feeds those points into GPT along with a suggested overall structure, GPT writes review.

• Human edits as if the review was written by another human.

The point here is that yes, the human had the ideas and finalized the product, but GPT did the writing. On the flip side, if all the human did was write about ideas GPT had, that would be a NO, because GPT didn't write anything.

This market only counts book reviews posted as part of the contest. If the selection mechanism is modified in a way such that the nominees are posted elsewhere instead, this market will be modified to be about those postings. (If enough changes are made to the selection process that figuring out how this market should be modified is confusing or controversial, I'll N/A it, of course.)