> Here we analyze 14 million messages spreading 400 thousand articles on Twitter during ten months in 2016 and 2017. We find evidence that social bots played a disproportionate role in spreading articles from low-credibility sources. Bots amplify such content in the early spreading moments, before an article goes viral.
I would also be interested in finding more studies that relate to what percentage of scientific research Americans thoughtlessly write off as "fiction" and the corresponding media diets.
- Inability to be replicated, not even in principle.
- Using definitions of "bot" that captures verified humans like e.g. the POTUS.
- Ignoring false positives in the data set or not even trying to measure them at all.
- Using opaque ML models trained off an irrelevant data set which (when measured by outsiders) have 50%+ FP rates.
- Logic errors and difficulty reasoning about causality.
- Mis-statements about what their own data actually shows.
- Refusal to acknowledge or address any of these problems.
There's a paper that thoroughly debunks this field here:
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3814191
and it's also available in talk form here:
https://archive.org/details/hopeconf2020/20200726_2000_Peopl...
and if you're in a hurry I summarized the paper here:
https://blog.plan99.net/fake-science-part-ii-bots-that-are-n...
NB: Some of my own meta-research into these papers is cited in their paper and talk.
Edit: I took a quick look at the paper you linked and it has many of the same problems. Amongst other things it explicitly classifies satire as "digital misinformation". One of the authors is Flammini who is well known for writing nonsensical pseudo-scientific papers in this space, in fact, he's mentioned in the talk linked above.
What evidence do you have that Twitter is lying about this? Why would they harm their brand by saying that they're infested with bot networks when they actually aren't?
1. https://www.nbcnews.com/tech/tech-news/exclusive-twitter-pul...