Om Malik blogs about censorship being a core goal in Chinese AI development:

Pete’s assessment suggests that Chinese authorities have shifted their focus from simply blocking information to creating sophisticated models that provide alternative narratives. You can’t filter these models, for they have finessed and developed evasion techniques and thus will be offering alternative sanitized versions of history that align with government narratives. This represents a fundamental shift from previous censorship approaches.

The sinking feeling in my gut tells me it won’t only be that government.