How to Regulate AI-Generated Content

The way in which all AI really is trained is using a database of some description. But when you are very closely censoring the information that's available, can you actually get enough data to train something successful? I do know that as you train AI's on publicly available data, the English language internet,. The availability of data out there in English is tremendously more than Chinese. Another thing is much of the Chinese internet, I'd say the diversity of things going on are actually offline. So for example, on WeChat and so on and so forth, is that really searchable or crawlable? And then third, actually in the recent draft regulations on AI-generated content that came out just