Do you actually think that will be the result? Why not the exact opposite? ChadGPT and the others are for all practical purposes trained to create content that is superficially appealing and plausible - i.e. perhaps not clickbait but a related longer-form phenomenon - without any underlying insight or connection to truth. That would make conversational AI even more of a time sink than today's clickbait. Why do you imagine it would turn out otherwise?