It seems more like the Chinese companies ar playing the dirty game, distilling through bot accounts, not letting real competition across their firewall.
So you are believing Anthropic's claim here, and it's not as if Anthropic didn't steal the data to train the model in the first place. I think the original sin doesn't give them any ability to complain.
- https://www.theguardian.com/technology/2025/sep/05/anthropic...