Bots, Facebook and cut-up poetry: An interview with artist and bot-maker Matthew Plummer-Fernandez

As if out of nowhere, bots are everywhere. Recent weeks have seen the very public meltdown of Microsoft’s Twitter AI, Tay – which turned into a genocidal racist in 24 hours – followed by announcements from Microsoft and Facebook that they would be creating bot platforms for businesses. Soon you’ll be able to order food or flowers by talking to a bot on Messenger, claims Mark Zuckerberg. Microsoft CEO Satya Nadella has even gone so far as to call bots “the new apps”.

But the art community has had an interest in bots that long precedes recent commercial projects. I spoke to game developer George Buckenham (@v21) about Twitter bots back in 2015, and last weekend I went to a bot takeover of Somerset House in London, run by arts organisation Abandon Normal Devices (AND) with support from Arts Council England and the European Commission. I even got the chance to make my own bot.

There I spoke to artist and exhibition curator Matthew Plummer-Fernandez about Facebook, Tay, and the playful, political works being made by the art-bot community.

(Above: Plummer-Fernandez (centre-right) and a crowd at the Art of Bots exhibition in Somerset House)

TM: There’s a lot of attention towards bots at the moment, what with Microsoft and Facebook both announcing their own bot platforms, not to mention the whole fuss around Tay. You’ve picked a great time for the show.

MPF: Six months ago, I had no idea all this was going to happen and take off down a commercial route. It seemed when I was talking about it six months ago [that] this was quite a microculture – an obscure folk art scene. A lot of these artists have been doing it for themselves, although some of their accounts have tens of thousands of followers. So it’s interesting to see bots now emerge as the new app, and to have bot marketplaces.

TM: Do you think that work like this takes on a different role when you have big businesses starting to use bots? Does it go from being a small scene to something more disruptive?

MPF: I think these artists really help to create some reference points and points of discussion. At the recent Bot Summit in the V&A, a lot of the artists were talking about Tay, the Microsoft bot, and critically unpacking what was going wrong with it – what is wrong with commercial bots in general. Is it because they have these gender stereotypes where you need a female assistant? Is it wrong for them to even pretend to be a human, or pretend to be an AI? These are things the bot community has given a lot of thought to.

(Above: Big Data Pawn Shop by Sam Levigne, Surya Mattu and Adam Harvey – a gift shop selling auto-generated items decorated with leaked NSA documents)

TM: I’ve talked to George Buckenham before, and he told me about the #botALLY hashtag. The community there is fascinating, with its discussions of bot ethics. Do you see that breaching into the mainstream now? Will those discussions about the ethics of bots become more common?

MPF

: Absolutely. The Tay controversy sparked a lot of mainstream newspapers to talk about bots. It’s nice that a lot of them have noticed a rich art practice that has interrogated this for a long time. I think the bot community have had a really enriching response to the rise of commercial bots.

TM: More generally, what is the appeal for the public? Why are bots popular?

MPF: A lot of them are designed to be entertaining. A lot of them construct jokes or humorous phrases that are designed to be engaging over a long period of time. Maybe that’s because a lot of the people that make bots come from a gaming background. They have the sensibility for making something that would be fun to play with over a long period of time. A lot of them are quite specific towards a target audience. For example, Darius Kazemi did a Harry Potter sorting-hat bot. By following it you’d be sorted into one of the Harry Potter houses with a personalised rhyme. That was instantly popular with a particular fanbase.

(Above: Matthew Plummer-Fernandez and Julien Deswaef’s Shiv Intiger, which pulls freely licensed 3D-printing models and randomly meshes them together.)

TM: I wonder if there’s an appeal in that people are aware it’s a bot doing it. With something like Magic Realism Bot, or Thinkpiece Bot, is there an appeal in people seeing a machine doing something that’s regarded as a human creative act?

MPF: I think the appeal is not so much in knowing it’s robotic, but more in knowing that it’s up to chance. There’s an element of serendipity in waiting to see what it will come up with next. There’s an understanding that this is a template and that there’s a list of possible nouns, concepts and phrases, and it’s jumbling them together. It’s a sort of cut-up poetry. So I don’t think it’s so much the sense that this is a robot, but that it’s a random form of poetry.

TM: Do you get a lot of poets making bots?

MPF: There are lots of people that come from an experimental writing background. Allison Parrish is one. In some cases they really have a skill of understanding the structure of grammar, and how you can quickly make recipes for making grammars.

(Above: Allison Parrish’s Everyword bot, which tweets out every word in the English language)

TM: Is there an overtly political use of bots?

MPF: There are certainly some overlaps. For example, I find Sam Lavigne’s work very entertaining, but it’s also very political. He’s made these generative edits of parliament debates and US news debates, where it automatically finds the keywords in the discussions and edits just to show those keywords being uttered again and again and again [@CSPANFive]. So I think there are different approaches. It’s possible to be critical, but also to be playful and poke fun at these systems.

(Above: An example of Sam Lavigne’s CSPAN 5 in action)

TM: That links to what you were saying about chance. There’s a kind of satire in how @CSPANFive reduces political debate to whatever happens to be the most-uttered phrase – it undermines the rhetoric. It’s interesting you mention this aspect of serendipity, because it seems to be quite different to how commercial bots are marketed – as calculating, intelligent, human-like voices.

MPF: The people that make bots in this community could make them smarter if they wanted to, but they just don’t want to go down that route. They don’t want to make complex systems that would baffle us with their complexity and their strange machine-learning behaviours. There’s something nice about the simplicity of being able to understand that simple processes can lead to lots of potential outcomes.

TM: I guess it’s also about highlighting the artifice of it all.

MPF: Absolutely.

TM: Do you see the art-bot community changing if commercial bots become a bigger thing?

MPF: I’m pretty sure they’ll get more attention around what they’re doing. It’s hard to say though. For me, it has been fun to intervene at a curatorial level and say these things can be performed and exhibited live. To at least show that bots aren’t necessarily just an online thing. Some of the bots – even Twitter bots – are bots that prompt us to do something physical. There’s one bot that literally gives you directions of where to walk. Even bots on Twitter exist physically on a server.

Read next: The art of the Twitter bot