As far as bot experience and doing bots actively for protest purposes, I guess I kind of accidentally did this with a bot called ClearCongress, which is @clearcongress. Not the best name, I think, but it was my attempt to try to respond to or deal with frustration I was feeling about political speech at the time.
This was right around the time of the federal government shutdown when Senator Ted Cruz and others were pushing the shutdown. The whole exchange was very frustrating to me and I wrote a letter to my local representative, and he wrote me back and it was this boilerplate of all this bullshit. So I was reading it and just mentally crossing off line after line of bullshit, so I decided to make a bot that does that for me. What it does is it goes into an API that lists members of Congress and their Twitter accounts, picks one at random, picks their most recent tweet, and it retweets it but replaces a certain percentage of all the characters of that tweet with black or gray rectangles so that the remaining amount of letters that you get to actually read is equivalent to the current level of Congress’ approval rating, which I get from The Huffington Post’s API.
It’s the kind of bot that the output of it isn’t all that interesting, like if you follow it you just see lots of rectangles and characters, but I think the idea of it is kind of interesting. I was interested in looking at the idea of bots as speech, or bots as being capable of creating free speech, and Mark Sample (I don’t think his name’s come up yet so I’ll go ahead and mention Mark Sample) had a really nice analysis of what he calls a “protest bot” versus an “activist bot.” He published this on Medium. It’s a way to talk about what kinds of things bots are capable of communicating, and I think protest and activism is one way to think about a certain kind of speech in a context like Twitter.
So, @clearcongress isn’t necessarily the best one but it’s mine. I also like the NSA PRISM bot which tweets out fake notices of the NSA noticing things, and when I look at all the different kinds of bots that I think are protest or activist, there’s a couple of themes that I tend to notice. So I wouldn’t correct any of what Mark Sample talks about as far as protest bots, but just to add a couple points of view, I think making something visible that’s not supposed to be visible is one aspect of bots in this vein. That’s also true of the broader field of so‐called “hacktivism.” And I think of @congressedits as doing a real good job of showing something that isn’t meant to be shown in a way, and that’s part of its energy or interest as a bot, and I think that’s a way to reverse that. So @clearcongress is the other side of that, taking something that’s supposed to be visible and then obscuring it in a way that is actually critiquing the context of the speech that’s supposed to be heard.
So I think that’s interesting and I think there’s a lot of interesting things at stake here when we look at political speech. Many of us are involved in academia and there’s been a few widely‐publicized cases of professors or other people being limited in their speech in social media, or being told that their speech in social media, their actions in social media disqualified them for jobs. This occurs to me as an opportunity for some further bot activism if we’re looking for ideas. I’ve been trying to think about ways to do this and I haven’t quite come up with anything yet. But it seems that there’s a lot at stake for some people trying to speak out in these contexts.
I don’t have much else to add about that, but I’m interested what other people think about ideas for speech as action and speech as refusal, if there’s opportunities to do that in a kind of public way. The nice thing about Twitter sometimes is that it can feel really small. And I think, to go off what Rob was saying, that when you choose your Twitter audience carefully (as we often do or sometimes it gets chosen for us that way) we have an opportunity to talk to each other or make refusal more focused or create locuses around actions or statements of refusal in ways that aren’t always available to us in other media or even other social media like Facebook. And so that’s why I think Twitter has a lot of interesting opportunities for activism.
So that’s a bit of rambling, I guess. It’s what I’ve been thinking about. I don’t have any new bots to show in that vein at the moment.
Darius Kazemi: So that wraps up the formal part of activist bots. We don't have a whole lot of time to discuss but if we have some questions and comments from people who are here or on IRC as well, I'm happy to spend five minutes on that and then regrettably we have to move on to the next section.
Joel[?]: I will really quickly say in line with what Rob was talking about. I haven't seen what his protection tool is or how it operates. I'm sure it's amazing. If you'd like access to something that you can make to help people you know who might have predictable criteria for who they might not want to be talking to or hearing from, I have been working on Block Together, and I've built one robot that indexes people who use particularly noxious hashtags and whose accounts were created after those hashtags began being used. It's a very blunt solution but it's been very helpful to a number of people. So if you want to give someone a link to an app where they can get an authorization and then you do the work of making a bot that maintains a list of who they can not hear from, that option is available and you can help people that way.
Darius: Excellent. I had a comment based on Ed's talk a little bit. It seems like once @congressedits became a known thing, then that invited people to sort of play with @congressedits. Like it almost instantly became a disinformation platform in addition to an information platform. And that just highlights one of the problems that I always have when I try to think up activist bots, is that you often end up designing things that cut both ways. I just wanted to point that out, really. I think that's interesting that I see that pattern over and over and over again.
Brett O'Connor: I have one comment. In regards to Rob talking about the anxiety around following people on Twitter and the social precedent Twitter sets through its UI and stuff, one thing I've always loved about the Twitter _ebooks bots is they'll take your Twitter account and kind of turn it into a Markov chain bot. And I think there's an interesting effect with a lot of people where seeing your own Twitter account distorted and broken in that way adds a little sense of liberation from how you're supposed to act or behave on Twitter, because the bot is just doing things you're not going to do. So you kind of see this alternate reality version of yourself and whatnot. I think that's another way that bots kind of challenge the socialization of Twitter and what they want us to do.
Darius: Thank you. Any other comments? Anybody else built an activist bot or anything, or protest bot?
Zach Whalen: Just one more thought. I think the idea of, as Darius said, the risk of these becoming a kind of performance and sort of just a performance for its own sake, which is what most bots are really, which is why we like them. But my take on it, I'm looking at or I'm interested in how bots are a kind of speech and that is itself kind of a fallback position, if you think about it. When I originally designed @clearcongress it actually just added the members of Congress. It just added their accounts, and that got it banned for a while and then I got it back in there. So now it's just performative, it just pretends to interact or pretends to retweet, and that is itself a way in which the API and what Twitter lets you do with it has decided that I can't do that in this case. I think the policy of not being able to massively reply to people, I think that makes sense that that's there for all kind of reasons related to harassment. But I essentially had this bot in a way that I felt was justified in harassing certain people in a very limited kind of way and I was denied that by Twitter. So I think it's interesting to see what Rob was talking about, ways to work with and against that API. That's really where a lot of these things butt heads, I guess.