LLMs are created through mass ingestion of data, and so, by definition, cannot be a clean room. They are trained on the things that they're tasked with copying.
I'm afraid, the companies in question, have been more than happy to ingest pirated material.
And the forefront models have intentionally been trained on quite a bit of reverse-engineering.
[0] https://en.wikipedia.org/wiki/List_of_commercial_video_games...
A clean-room implementation is going to be much, much more effort than a "decompilation". Even a decompilation is, I think, more than the hype is claiming. You can't just point at a game, say "decompile it and make it work on Windows" and get a result in one shot. It will take playtesting, cycling around, QA, and time to polish it up. But that will still itself be much easier than an actual clean room implementation. Let alone getting sued by the owner anyhow and having to prove it was clean room in a typical chaotic individual dev environment.