Read on Twitter
NPR is promoting this article again. Without access to the study we have no way of knowing how "bot" was estimated or measured, we simply have to go on the reputation and past research of this lab, which I dug into last night here: https://twitter.com/tinysubversions/status/1263675864568356864?s=19 https://twitter.com/nprpolitics/status/1263947771788308482
The short of it is: knowing what we know about the study, which is very little, it seems like these researchers have in the past used a very loose and nearly useless definition of "bot"
But even on top of that, use your brain: of the human people in your life, your friends and family and coworkers who you follow on Twitter... what percentage of their posts have been about the pandemic in one way or another for the last 2 months? For me it's easily above 50%.
COVID-19 is all anyone is talking about, and unless we posit that there are more bots than people out there on social media, there needs to be extremely good data to make a claim that half of all conversation about COVID-19 is from bots. The burden of proof is huge and not met.
Also worth looking at is this informal audit of a few "bots" that were identified by these researchers back in April, some of which are humans with faces and lives who post videos of themselves like, talking and living and stuff https://twitter.com/FlorianGallwitz/status/1254008280571170816
Also if you're interested in this you can check out my blog post on "The Bot Scare" which is not peer-reviewed but I try to cite lots of sources and make a decent argument that most of this kind of research is pretty flimsy.
https://tinysubversions.com/notes/the-bot-scare/
https://tinysubversions.com/notes/the-bot-scare/
@BobbyAllyn I would happily talk to you about this issue and help set the record straight, you are welcome to look up my name on Google Scholar if you need to see the wide range of citations of my (non academic) work on bots
This story continues to be widely circulated today, this time though MIT Technology Review. Again: this is based on a press release with no published data or methodology, and the bot detection work by this research lab in the past is questionable. See my tweets above for why. https://twitter.com/JonLemire/status/1264180035646296067
Here is my original thread responding to NPR's coverage of this press release without a study. https://twitter.com/tinysubversions/status/1263675864568356864?s=20
Extraordinary claims require extraordinary evidence, and this particular claim is accompanied by literally zero evidence. Instead, they are trading on the name of Carnegie Mellon University, saying "trust us, we're CMU scientists."
I'm very mad today at news outlets who fall for the "trust us" line and will publish an article about the claims of scientists without linking to a published paper or at least a preview of the paper.
One irony is I *want* to believe that 1) there is a COVID bot problem on Twitter and 2) the bots have a measurable effect on the discourse. It would be a relief, in a way, to know that people are being manipulated into behaving poorly. But so far, no evidence either way.
Just scrolling through the replies and quote tweets of the NPR and MIT Tech Review articles I have found hundreds of discussions and comments mostly supporting these articles, which are themselves unsupported information!
I think I am going to have to do a more formal analysis of how this misinformation *about* misinformation makes the rounds and affects the conversation. I promise I will publish my data before making any press releases.
Here's Hillary Clinton posting the NPR coverage of the article. Again: no study, no data, just claims made by people who have not done very good research in this area in the past. Makes for a great headline, though. https://twitter.com/HillaryClinton/status/1264225262885056513
