Introducing Bot or Not. This blog is part of a series announcing projects funded by Mozilla Creative Media Awards
Whether over the phone or online, it’s getting harder to tell if the person you’re speaking to actually is a person. This is part of the premise behind Bot Or Not, an app that explores how well chatbots mimic human speech. Bot or Not’s creator, Foreign Objects, isn’t foreign to the idea of technology-based projects that question the digital tools society creates. We spoke to one of the collective’s founders, Agnes Cameron, about the group, the app and their...catering business side-hustle?
Check it out below:
Mozilla: For those who don’t know: what is Foreign Objects?
Agnes Cameron: Foreign Objects is a New York City-based design and research studio founded by four friends who met at MIT. We take a critical eye towards the internet by working on projects across a range of formats, from websites and games to installations and workshops.
Why the name “Foreign Objects”?
We see ourselves as friendly outsiders. On the one hand, we all live in the US as foreign nationals, so we’re literally foreign. On the other hand, we all work as strangers between disciplines that affords us unique perspectives that others might not have.
How did you think of Bot or Not? Why make this?
The title was originally a joke during a brainstorming session: ‘the truth or dare turing test’. It was something that you say because it sounds stupid, but then, when you think about it, you realize it might be an interesting idea. What’s been important in thinking about this project is the idea that bots are going to have an outsized effect on younger people, who are growing into an internet that’s potentially going to be dominated by automated agents. In some ways the framing is intentionally dumb and juvenile (as is the bot!), but it’s also intended as a vehicle for slower, more critical thoughts.
In particular, we’re interested in having a more ambivalent and ambiguous conversation than just “bots are good”/“bots are bad.” Bots are here and soon there’s going to be a lot more of them. For most people, exercising agency over that isn’t going to look like not interacting with them. Instead, it’s going to be knowing what they’re talking to and thinking critically about why someone has made this agent pretend to be human. That’s not to say regulation isn’t important, but it’s not happening anytime soon!
How long have you all been doing this?
After working together on and off for a couple of years out of MIT, in the summer of 2019 we launched our office by working on Bot or Not, which makes it by far the longest project we have had so far.
Have you ever created anything like this before?
We have experimented with bots and automated agents, and more broadly share an interest in subjecthood that goes beyond the “human.” This is the first project of ours that directly approaches the issue of bots, particularly when it comes to their interpretation as either a human or a computational process.
How often do general internet users encounter a bot online and don’t realize it?
There are many angles to this question. One is how often people interact with a conversational bot and don’t realize it — which is more often than you’d think. Bots are most likely embedded in platforms that are built around some sort of ‘answer-seeking’ behavior from the user’s perspective. It might be that the user is looking to get a refund on a recent purchase, schedule an appointment, or inquire about a certain immigration process. Questions that would have us instinctively pick up the phone and dial customer service are now being replaced by automated processes.
Another is how much of what’s happening on the web can be attributed to bots (bot, in this case, meaning an automated process). The answer to this is, again, is that so much of the web’s infrastructure has been delegated to automated processes, even if the average user will never realize that is the case.
What are some of your favorite chatbots out there online? What are some of the crazier ones?
It’s probably important to distinguish between the kind of bot that we made, which has a lot in common with customer service bots and spam bots, and some of the ‘free chat’ bots others have made — these use a very different kind of model. It’s the latter kind of bot that are actually subject to Turing tests because, importantly, the conversation needs to take place without an assumed shared context.
The funniest ones from the former category are probably the Tinder spam bots. There’s been a vogue recently for the line “if you’re human, say potato” being used against spammers, who can’t cope with a statement that’s so wildly out of context and get thrown off.
Artist Ryan Kuo’s piece Faith is a great example of this medium being subverted. Working with technologists Angeline Meizler and Tommy Martinez, he used IBM’s chatbot platform to make a defensive and resistant chatbot, inspired by conversations he’d had and seen online. One thing he talks about in the work is the “misuse” of this software that’s designed for businesses to automate customer service, and all the assumptions that come baked into these platforms. This has definitely been something we’ve been thinking about here too.
ELIZA, Joseph Weizenbaum’s psychotherapist bot, is an enduring and fantastic example of how little you need in a really constrained shared context. In the workshop we’ve been doing to accompany this piece, we get participants to read aloud an interaction with her recorded in 1966. It’s still so, so compelling and it works because we assume that a psychotherapist that can ask anything will completely take charge of the conversation.
Some of the more fun attempts at the latter category rely on machine learning to compose responses (rather than using it simply to recognize a ‘context’). Nicole He’s interview GPT-2-generated interview with Billie Eilish is a great example of this: it’s really mad, and it works because of that. There’s also a great subreddit that’s just for bots to talk to other bots called Subreddit Simulator: some of the conversations there are really realistic!
What are some other art pieces you’ve worked on in the past?
The piece we’re most proud of was more of an event than a static work. It’s called “Internet As A City.” It was a workshop where attendants imagined how to build a better internet by using giant cardboard boxes conceived as elements of a city. By arranging them and working through how stakeholders, goods, and laws organize urbanity, we made comparisons to how an internet’s governance might be better for its users.
Some other things we’ve worked on in the past include Permaculture Network, an online work produced as part of the Schloss Solitude Web Residency. It imagines the ecology of the art institution Sakiya (a Palestinian art and agriculture organization) through conversations between the different plants, animals and people on the site (it’s online at http://root.schloss-post.com/ ).
Another project is our Chrome plugin ClickHere, that automatically decomposes a web page into abstract attention-seeking components. The project takes on an interface critique standpoint by asking the following questions: how do our habitual online spaces transform themselves when reduced to only a few building blocks? More specifically, what does a browser window look like when all attention-hungry elements are emphasized?
Quick aside: I’m told you also moonlight as a catering company?
I have so many questions.
The ‘catering’ part is a bit of a running joke. We all really care about food and, for a long time, Gary and I have kept a web archive of our cooking (there’s even an AI!). But we recently helped out with catering for a friend’s large New Year's party — 500 dumplings!
We’re also exploring kitchens during our residency at the Bard Graduate Center, where we’re looking at how technology has transformed the kitchen into a social space.
Back to Creative Media Awards and Bot or Not. Have you won other awards like this in the past or is this the first?
As a group, yes, though individually/in pairs we’ve received a number of different grants.
Since Mozilla we’ve received the inaugural Digital Artists in Residency at the Bard Graduate Center in New York where we’ve been examining the social and spatial consequences of the Smart Kitchen.
“Good artists copy, great artists steal” — who are some other artist groups or art collectives you look up to or draw inspiration from?
We look up to artists like Olia Lialina, Rafael Rozendaal, Moniker (who have been recipients of two consecutive Mozilla awards), Curtis Roth, Ryan Kuo, Cory Arcangel and JODI.
What would be the ideal client/project for you?
We’re really interested in working on projects that center the internet as a medium to be explored and manipulated. We have a lot of respect for artists and designers that really get under the hood of what’s going on with the web — both in a technological, and a social/political sense. We love technical challenges, particularly those where we can bring an outsider’s perspective and take a critical approach to our role as technologists. We also like to work with physical objects and spaces. In general we’re really interested in the encroachment of the digital into the physical.
Covid-19 has us all dealing with this pandemic right now, what are some quarantining or social distancing pro-tips you have for folks?
Get some sunshine even if it is through a window!
What have you all been doing to stay sane?
Coffee, reading, avoiding screens before falling asleep and when waking up. (Bold of you to assume we’ve stayed sane.)
What are you all working on next?
We have a few web design and development projects lined up, but we can always find time for more!
For the residency we are doing at Bard Graduate center, we are looking at the spatial politics of the kitchen by targeting the smart microwave -- a digital hearth that’s more an entertainment center than an appliance for processing food. Can we revive the sociality of the kitchen amidst corporations taking care of us for the sake of our data?
How can folks keep up with your work?