I'm a tech optimist and a strong proponent of AI tools. But I also admit that the Chinese internet is polluted with AI-generated content, and it's going to get worse. If you want to blame the AI for being too easy to use, blame the people who are too good at drilling traffic. With the support of the advanced technology of artificial intelligence large language models, the content farms of the year are outputting garbage content on a large scale and in an organized manner. Therefore, being able to recognize AI manuscript washing will be a skill, and cultivating critical thinking and maintaining independent thinking is a basic literacy. WeChat***I can brush a lot of 10w+ of AI every day, in order to avoid bad inducement, I won't give examples. There are also many people in the WeChat circle of friends who are doing similar AI explosive article writing. WeChat's approach to this is to require authors to choose to label the content generated by AI when creating
I don't know what the impact of labeling will be, but none of the AI content numbers I've seen will be labeled. If the annotation of artificial intelligence generation will limit the flow or even cancel the revenue, obviously no one will take the initiative to voluntarily mark it, after all, they concoct these contents to make money. On the other hand, even if users want to complain, there are few channels. In the complaint page of the WeChat *** article, there is no reporting option for AI-generated content, and the definition of abuse of originality or plagiarism cannot be included. 
From the perspective of the platform, if there are no mandatory regulations, there will not be much enthusiasm to actively control such spam content, because even if it is a content farm, it is also traffic, and there are people who watch it, and it does not affect the point of advertising. But if it is the other way around, if someone uses some kind of technical means to pick up the wool of advertising clicks, it will definitely be controlled soon. In this comparison, Zhihu is actually not the most widespread place for Chinese Internet AI manuscripts, and Zhihu is also trying to govern AI-generated content. But really, it's not that the quality of Zhihu users is higher than that of other platforms, it's just that they don't make money here, and AI farms don't even bother to do it here. Of course, there is still a lot of AI content on Zhihu, especially under the red envelope questions and the daily hot list, you can brush a lot. Some of the AI eyes will be dealt with as long as they report it; However, as long as you change a few words or change the tone, it is difficult to identify the report. Because I use AI intensively, as long as it's not artificially refined, no matter how much I tune the prompt, it's still an AI. But the success rate of reporting is only half and half, and sometimes it's quite annoying to see similar content. 
The following are the reports that failed:
It's not that I tear my umbrella or smash the pot, but it's really pointless. It's annoying to watch too much, and I don't want to know it in the future. A few more idle words. Large language models, which first became popular a year ago, have focused on content generation to plug-ins, to agents, and then to RAG and the recent AI search. Developers are actually trying to reduce or eliminate the illusion effects inherent in large models. From Newbing to Perplexity, one of the benefits of AI search is that it is annotated to verify the source. As mentioned before, Zhihu has a high proportion in Chinese Internet search engines. But if it's all AI content in the future, then no amount of RAG and source verification will work, and it is equivalent to poisoning from the source. At present, there is no good way for AI companies, and what they can do is probably labeling, but labeling from the source. I was pessimistic about this issue last year, and a year has become even more pessimistic...Maybe one day, AI content will flood the country to the point where only human-generated content will be additionally labeled as Handmade:(