#109 - Samantha Bradshaw

**Craig Cannon** (0:00)
Hey, how's it going? This is Craig Cannon, and you're listening to Y Combinator's podcast. Today's episode is with Samantha Bradshaw. Samantha is a researcher at the Computational Propaganda Project, and a doctoral candidate at the Oxford Internet Institute. She's been tracking the phenomenon of political manipulation through social media. You can find Samantha on Twitter at sbradshaww. All right, here we go.
So I wanna talk about two separate things.
So one is the responsibilities of these platforms and the responsibilities of governments. And then the other thing is bots, because I think a lot of people hear the word bots, but don't necessarily know what that means. So maybe we should start there, and you can kind of contextualize what bots are actually doing right now. And then we can talk about how these corporations are interacting with their users and with governments. So how would you define a bot?

**Samantha Bradshaw** (0:52)
Yeah, so I mean, a bot is essentially just a script or a piece of code.
And I mean, bots can do a whole bunch of different kinds of things. Let me start here. So a bot, for example, could be a web scraper, like Google's search engine is essentially a bot. And what it does is it goes out and it crawls the internet in an automated way and then stores information about different pages. And so like, that's a really good bot that we need for the internet, because if Google wasn't able to scrape and crawl pages in an automated way, it wouldn't be able to index them and we wouldn't have Google search.
Other bots, and I think the bots that are being talked a lot about in the media and around all these questions about Russian foreign interference in elections and manipulating the media, these are bots that people design and they're designed to mimic human behavior in an automated way. So they might plug in to Twitter or into Facebook and then they might like, share, retweet a whole bunch of different stories very quickly to give something a sense of popularity, like, ooh, lots of people are really engaging with this content. They might follow people just to, again, sort of amplify that popularity, maybe around a particular person rather than an idea. Some of the more sophisticated bots might actually interact with real people. So they blend like chat bot technology.
And we all have been to those customer service pages where the thing pops up at the bottom and says, hi, can I help you today? That's not actually a real person. That's another kind of bot, but it uses a lot of natural language to interact with people. And so that kind of technology can be plugged into some of the bots that we're seeing on social media platforms to actually respond to users in comment threads or post like a little comment about a story that's being shared.

**Craig Cannon** (2:54)
Yeah, and when did it become clear that governmental agencies or non-governmental agencies were using bots to manipulate thoughts around politics?

**Samantha Bradshaw** (3:04)
Yeah, so I think this phenomenon really came into media and public attention during the 2016 election.
But from some of the research that I've done here at the OII and on the Computational Propaganda Project, we know that these techniques have been experimented with by governments for a long time. We've had evidence of them going back to, you know, even the early days of social media platforms and using these kinds of tools to shape discussions mainly at home and in more authoritarian regime contexts, as, you know, another tool of social control.

**Craig Cannon** (3:44)
And so now that you've been studying this for two years, so you came on in 2016?

**Samantha Bradshaw** (3:49)
Yeah.

**Craig Cannon** (3:50)
And so then pre-election, you started doing this research and now there's a lot more attention paid to it. How have bot tactics changed?

**Samantha Bradshaw** (4:00)
So it's been fascinating to study this from the, you know, sort of inception of attention to this issue, because even when I started this research agenda, you know, the whole focus was on Russian interference in the US election. But, you know, I just had this inkling that it was a much broader phenomenon. And, you know, there's more, all kinds of these digital techniques being used that, you know, make use of automation and algorithms and trying to game them to manipulate public opinion.
And so, in terms of how I've seen this change over time, we're definitely seeing a lot of new entrants starting to experiment with this kind of technology.
And, you know, in particular around elections, usually these new entrants might experiment with some of the more crude bots. So things that might just like share and retweet certain kinds of stories or follow a politician. So they don't really engage with users very much. It's usually pretty easy to tell that these are bot accounts.