XD
Didn't work for me. It played perfect theory until move ten, when it tried playing illegal moves. I think it just uses opening theory from online until it runs out of it.
I always considered ChatGPT as what would come out if Google and Wikipedia had a child. As it's based on reinforcement learning, I could imagine it having some kind of Leela Zero built in.
The only time I asked it something related to chess it made several mistakes in the answer. At least it apologized when I pointed out one of the mistakes.

XD