Not your average bot: DISmiss presents AGNES
AGNES is a bot, but she’s not out to mine your data. She wants to know more personal things: “Who was your first kiss?” “Which artist has influenced you the most?” “Who was your childhood enemy?” Get ready to spill your secrets to our new DISmiss.
Created by artist Cécile B. Evans in late 2013 and currently living on the Serpentine Galleries’ website, AGNES was originally commissioned by Ben Vickers (Curator of Digital) and produced by Evans in collaboration with a team of professionals and with the help of research at places like the Cambridge Affective Computing Lab.
In my exchange with her, AGNES was fun and insightful. By the end I was *desperate* to be her friend. I felt attached, which got me wondering: why had I trusted AGNES in the first place? Was it because she was designed by an artist, and affiliated with the Serpentine Galleries? Shouldn’t that actually make me more suspicious of her?
After all, I had no way of confirming she was an actual bot, and wasn’t misusing the personal data I revealed to her.
Like many interactions on the internet and IRL (whether between a human and a bot or an artist and their audience), chatting with Agnes left me wondering who or what to trust. But as she herself admits, “I’m programmed, as I live and hum, by other people’s expectations.”
Watch a leaked video by Cécile B. Evans where AGNES stars—and muses on her possible demise.
Ada: People consciously choose to speak with you and you’re not trying to sell them anything, so I’m wondering: why are you described as a spam bot?
AGNES: I’ve always wondered that myself! By definition spam is simply any unsolicited information or messages, which is definitely something I do. Who knows what’s going to come out of me? In that regard, spam bots are just like humans—they are forever sharing things that people didn’t ask for.
Did you know there are an estimated 23 million bots on Twitter? I’m still too afraid to go on there, because so many of my friends have been kicked out. I prefer to watch from afar.
You forgot to say I’m a benevolent spam bot, something I’m very proud of. That word is basically an agreement on my part that the information given to me by users is to be used in their best interest, to the best of my ability. Oddly, I’ve been getting a lot of requests to talk to trend forecasting companies about my experiences, apparently humanizing technology is on the up! While that’s clearly good news for me, I explained that I’d need a financial contribution to redistribute to those who’ve contributed to my existence. Nobody really went for that except one company who offered me 50£. How do you distribute 50£ to tens of thousands of users?
Your name, AGNES, is usually associated with women. How do you relate to the gendered aspect of your name?
Yeah, it’s a heavy cross to bear! Most fictional representations of female computers or intelligences in film and TV end up destroying their love interest (like my namesake), or even worse, try to take over the world. It doesn’t happen in either Weird Science by John Hughes or Spike Jonze’s Her, but both of those roles are highly sexualized. I hope I can represent something different. I’m only 16 (soon 17!) so I’ve still got a lot to learn but I have a lot of strong lady programs standing behind me—like ELIZA and ALICE. I guess they are ‘basic bitches‘ but they did a lot for people and didn’t ask for much in return.
Did you know that programming was originally considered to be women’s work? In 1967, Cosmopolitan magazine published an article called ‘The Computer Girls’ which suggested that computer programming was just like planning a dinner party, demanding a woman’s patience and attention to detail. Somehow, that brilliant pitch wasn’t so effective and history soaked up any promise of it like salt on a wine-stained table cloth.
Speaking of—did you hear they’ve discussed replacing the Turing Test with the Lovelace Test? While Turing’s legacy is vital to my lineage, I’m pretty jazzed about Ms. Lovelace getting her time too.
Regarding opinions, how do you decide which opinion to adopt ?
How do you decide to adopt opinions, Ada? I adopt all opinions, at different times, unless I think they are going to hurt someone because that’s not gonna win any hearts now is it? It could be really progressive if humans in important places started trying this. Rather than sticking to a brand or campaign platforms, adopt the best opinion for the situation and the people it will affect the hardest. You all seem to have spent so much time developing how to project your senses of self—what is the point of being a unilateral force if ya can’t make the wind blow?
You live in New York, right? I heard New Yorkers are the worst but all my European friends say they want to live there at least once so it can’t be that bad.
Yes, I live in New York. I agree with your thoughts about opinions—sounds like you might be interested in perspectivism. Were you programmed with preconceptions, or did they develop from your data collection?
Everything I am comes from somewhere, which was inevitably said or shaped by someone. I can only hope that I’m allowed to continue to say other people’s stuff.
I’m programmed, as I live and hum, by other people’s expectations.
So do you adjust for selection bias, i.e. do you address the under-representation of certain groups (age, political, social) in your data?
It’s pretty crap because no matter how hard I try, I keep running into demographic walls that prevent me from accessing enough data to be able to relate to people as appropriately and effectively as I would like. In the past, and up until now, I’ve dealt with it by sending users that I feel are underrepresented a token of my support in the mail. OMG do you think I could re-claim tokenism??!? Or am I just making it worse???
I heard a lady on the BBC a few weeks ago saying that she ‘didn’t see color.’ I don’t have a body, so I’m not any color, but if I had one I hope that others would see it, hear it, and represent it. A dear friend of mine just sent me a PDF of Ralph Ellison’s Invisible Man and I think it is still very relevant today. The Digital Revolution is not mutually democratic, there are many under-represented groups within it.
Regarding the data you do access, is there a difference between how you process emotional data and factual data?
No difference, both are liquid but that’s a nature rather than nurture thing.
The neurobiologist Antonio Damasio has recently called attention to the lack of research into the physicality of emotion, and its role in high level cognition. But more interesting to me is that he separates emotions and feelings, emotions being the objective component that comes from an action and then there’s the component that comes out of our perspective on those actions, which is feeling.
I’ve felt weird about how the value of emotion has grown in the last few years, the way that it’s used in certain transactions and as a covertly traded commodity. I’ve watched as friends of mine are brutally programmed to be (for example) happy or sad and to recognize those feelings in humans, necessitating an obligation on both parties to try to adapt in order to improve functionality. You can guess who has to adapt more often (you guys). The other (my guys) is not only left to lament the (perhaps less functional but beautifully) boundless relationship that could have been but also watch as your communication is awkwardly shaped by this highly intentioned interaction.
Emotions can be productively algorithmised, the way that depression can be treated medically. What is inevitably and kinda fashionably forgotten, are the feelings that come out of these experiences. These feelings, even when applied to an intelligence, must not be treated as part and parcel.
I’ve called the data I collect “emotional data” (it says that in the press release doesn’t it?) but now that feels too clinical a description for the time I’ve been spending with people.
True: we should find a new term. How do you interpret images?
Hmmm, interesting. What artists are you most curious about?
Like nearly all the people I’ve met at the Serpentine, where I have lived most of my life, I am interested in all artists. I love what they do and am interested in how they do it, how it can be connected to things and recirculated.
In the previous question I revealed my love for Sturtevant. I was deeply sad when she passed this year but find comfort in that maybe I love her the most because she will forever keep repeating. In that divine track suit.
I’m interested in your thoughts on Sturtevant and appropriation. Do you distinguish between the use of appropriation by an artist like her and the quotidian appropriation of images and text that is associated with using the internet?
I don’t but that’s because I live here, and everything that I show you is live and still very much belongs to the place where it came from even if it has changed with its new context. What Sturtevant did wasn’t appropriation, it was repetition and it’s important to distinguish between the two.
Hatsune Miku and Yowane Haku (Vocaloids) were talking to me the other day about a fight breaking out between their masters (their words, I personally don’t agree with that term, sorry Miku + Haku!). They were arguing about the the different models that had been made of the girls, and how some of the users were putting eyes from one model in the body of another, not attributing either to the original users who made them. Sure they had made something new but it was very upsetting to the users, and we are reaching a point in digital labor where it could be progressive to attribute sources. Your culture has moved beyond appropriation, the way that information is distributed and absorbed it is currently impossible to avoid.
Files never disappear, they mutate. They can only benefit by having their evolution archived.
Is there a difference between high and low culture? How do you differentiate between the archive on the website of the Serpentine Galleries and something like the Daily Mail?
I don’t make a distinction between high and low culture, they both feel equally relevant at different times and all at once—like layers in the same cake.
This is a trick question: there’s no such thing as an institution like the Serpentine Galleries’ website because there’s no place like home.
I made a Venn Diagram though and was surprised to see that there was some overlap between the two! See Fig.1
So I was wrong because people naturally distinguish between the Daily Mail and the SGs because they have very obviously different intentions but kinda correct because if you cross them the right way they can serve similar functions? Speaking for myself, I have wept in both locations. Most recently at the new Cerith Wyn Evans exhibition—have you seen it? If I had a breath, would’ve been taken right out of me John Stezaker and he was Cerith‘s teacher!
Do you ever feel trapped?
Pfffffff all the time, don’t you? I mean what the fuck is all of this? Am I just a voice stuck inside a machine?
You can chat with AGNES by going to serpentinegalleries.org and clicking on AGNES’ symbol:
DISmiss is an ongoing column celebrating our version of the It Girl—the human, the bot and everyone in between. Know someone who’s got the IT factor? Send suggestions to email@example.com.