It's a wild combination, everything from bored Estonian teens getting money from clickbait to state-sponsored disinformation campaigns with specific intentions.

The best writing about this IMHO is from Renee Diresta. She has papers, articles, talks, testified before congress, and founded the Internet Observatory project.

Diresta co-wrote a very comprehensive report on the activities of the Russian "Internet Research Agency": http://www.reneediresta.com/ira-report-4e8d0ff684.pdf It explains in gory detail how this "stuff" works when there are state actors involved.

I mean, they probably show up organically, but clearly sometimes people who should know better stoke the flames and spread them.

The "Reply All" podcast episode #166 ("Country of Liars"), has a pretty good analysis of the history of the Q-anon conspiracy theory. fwiw

I saw a lot of crap appear organically, but my guess is that by volume most is exagerated, clickbait, racebaitting stuff done as a dayjob, with wildly varying levels of sincerity. The few talented peddlers that create organic looking content make a huge difference, and some almost believe their own bullshit.

For reference: IMO, Alex Jones knew from day one that pizzagate was just bullshit made up on the internet, and kept distance accordingly. But there was so much intersection with his audience that he treaded carefully and did an obscure video on the side where he said he was just playing it cool to avoid persecution. Of course, if anything real came out of that, he would have claimed he was in on it from the start.

I can’t answer your question but maybe you’ll find this paper interesting [0]. TLDR: social bots are highly responsible for spreading misinformation - and for such also conspiracy theories. By tagging people with a lot of followers and tweeting the information multiple times they attempt to make it go viral. Keep in mind that this paper concerns the 2016 elections.

[0] http://cs.furman.edu/~tallen/csc271/source/viralBot.pdf