Comment by anigbrowl

This seems like an especially clumsy attempt at chatbot influencing/search engine optimization, bu tit's symptomatic of a much winder problem. Regardless of the underlying issue, there's a non-profit-industrial complex that has turned think tanks and opinion formation into a science, and a whole PR industry masquerading as news which is worse than the 'yellow journalism' of generations past.

Disinformation and cognitive spam are all forms of information pollution, and our information environment has been massively degraded over the last few of decades - both by greedy and short-sighted opportunists, and and by cynical sociopaths who know exactly what they're doing. I think AI has great potential to be a supplement to media literacy and to effectively categorize and filter news-assertive content based on its timeliness, factual density, and presence of absence of emotive and rhetorical manipulation. But if we're not careful we're going to see the environment being wantonly polluted for the express product of shaping model outputs.

It remains to be seen whether objective measurement techniques from the NLP field combined with pruning and other model refinement techniques can limit or reverse propaganda undertakings like this.