Zeynep Tufekci: uncertainty

Transcript context

…It keeps on showing me chess videos. Maybe the Sicilian defense pops up. But there is an alternate literature. Munger and Phillips, Ledwich and Zaitsev— those have written papers in the last three years saying no evidence that YouTube radicalizes. Maybe that’s agnostic. When I read you on lab-leak, it’s all about agnosticism, expected value, wisdom, multiple perspectives, but when I read you on social media, YouTube, it seems very hardcore. “Here’s the right answer. It’s bad for people. It radicalizes them. It’s all about the ad.” It’s like these two Zeyneps, and why not apply the lab-leak approach to social media? No, no, no, I’m going to push back. Well, here’s the thing, though. I’m going to push back a little bit because my book is all about how great social media is, in some sense, for dissidents as well, but it’s not an unalloyed good. It’s complicated. I’m from Turkey. It would be hard for me to deny that Facebook is now the public sphere there, and Twitter. That’s where you break a lot of the censorship. I’m acutely aware of a lot of those things. But in terms of the algorithm pulling you to the edges, I do think it’s still true. It doesn’t really mean it’s pulling you to only one side. That’s the misunderstanding. One of the few pages that does really great on Facebook on the left is something called Occupy Democrats. It’s just nonstop outrage and barely tied to truth, so it’s not just a right-wing thing. The fact that it is very friendly to authoritarians for some reason, but the fact that it was more right in the United States, is more about the United States’ political structure. My criticism isn’t that it’s great for us to have these large platforms, in some sense, that greatly expanded speech, but the way they keep people on the site by keeping them interested is a bit like keeping people in a cafeteria by giving them constant chips and ice cream. I like chips. I like ice cream, but I wouldn’t want a cafeteria whose business model is to keep me sitting in that table as long as possible. Because we know this. We know in cities where there’s the infrastructure for walking, people walk. When you change the menu in cafeterias, people eat different things. When you put stuff in different places, people watch different things. I think the papers that you cite — one of the problems here is, it’s a little bit like China in that we have no data. The YouTube stuff that we have tried to do — we’re literally looking at very, very, very broad data because what we need is YouTube’s internal data, perhaps linked to a panel. Think not the GSS [General Social Survey], but something like a couple-of-thousand-people panel that is linked to data. It’s 20 years into the internet explosion, and we don’t have a single good study. Most of the data is locked up in the companies, and there’s not a good research thing. I’m agnostic in the sense that I don’t really know exactly the extent to which the algorithm works, but that’s similar to lab leak, where I don’t know what to say because whatever data there is, it’s in China, and we don’t have access to it. In all of these dialogues, there’s a middle segment called underrated versus overrated. I toss out some names, ideas. You tell me if you think they’re underrated or overrated.…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.