It depends on how valuable it is to spam / fake engagement. I can see wealthy bad actors expending a lot of capital to influence public opinion. These AI texts are extremely persuasive and convincing.

Converesely, those with the biggest secrets to spill tend to be lone actors not backed by the state. For example rebels reporting on the mass protests in China could easily be drowned out by AI-driven CCP propoganda.

One could establish a web of trust, and content could be vetted by digital signature schemes. But A) Finding trusted people in the first place can be tricky, and B) We have to coexist with normies, and normies won't bother to do this.