THE BOT HAS AN AGENDA: An Interview With Christopher Fahey

THE BOT HAS AN AGENDA:
An Interview With Christopher Fahey
[graphpaper.com / askROM.com]
w/ Eryk Salvaggio

++++


Eryk Salvaggio: Can you describe your first work of art?

Christopher Fahey: I assume you mean the first work of art I ever
*made*, not the first one I ever saw. I've always been making and
drawing things, and I've always been coming up with interesting and
totally impractical ideas. My childhood was filled with model airplanes,
home-made computer games, science experiments, and lots of comic book
drawings. I just made lots of what I thought was "cool stuff", and
didn't really distinguish between art, experimentation, and hobbies.
Throughout my childhood I've spent a lot of time in the Marcel Duchamp
gallery in the Philadelphia Museum of Art, probably one day of almost
every year of my pre-adult life. But only when I started imagining
myself as an artist was I able to start imagining that I too might be
able to make those kinds mysterious and idiosyncratic objects. When I
was 16 or so, I saw a major traveling exhibit of the works of Jonathan
Borofsky and I was hooked.

ES: Why did that show strike you so much?

CF: It was at the Philadelphia Museum of Art. The show was like a big
playground. There was a ping pong table that visitors were allowed to
play. There were gigantic kinetic sculptures doing these absurd menial
motions like push ups and hammering the air. Some spoke, muttering
"chatter chatter chatter". Borofsky had a personal kind of meditation
where he counts from one to infinity by writing numbers on graphpaper.
He'd spend some time every day writing numbers… 1,456,293…
1,456,294… 1,456,295. The show was my first introduction to a lot of
art concepts that I had never known about before: Interactivity,
multimedia, installation, stylistic and conceptual diversity, the
significance of the artist's life as a kind of overarching meta-artwork.
Even my love for graphpaper. The show was fucking hilarious, too. I
learned that artists were "allowed" to do a lot more than I had ever
thought they were allowed to do.



ES: So what's the project you're working on now?

CF: I'm working on an artificial intelligence personality that will be
an alternative interface to the Rhizome artbase. Using a simple chat
based interface, users will be able to either explore my character's
personality or explore the Rhizome artbase.

ES: I noticed you are going to do the Rhizome bot through AIM, right?

CF: I am totally addicted to AIM.

ES: AIM is interesting to me because we had phones which replaced the
telegraph and gave you the human voice, and now we're using phone lines
to go back to a telegraph style of communication with all this chat
software.

CF: Instant Messaging is funny - it's like a phone conversation where
you are not socially required to reply to your friend instantly. You can
wait a while, or even not answer it at all. It's fascinating to me how
new technologies invent their own models of social behavior. Instant
messaging is something that nobody thought would work - it inhabits this
middle ground between email and the telephone. But almost anyone who
uses it falls in love with it because it creates a social model that is,
in fact, quite convenient.

The general technology of Instant Messaging is sometimes called
"Presence Management" or something, because the communication itself is
not what's important - what's important is how it enables two
communicating parties to communicate in a way appropriate to the level
of availability of each participant. If I'm at my desk, you can IM me
and talk to me right now. If I'm not at my desk, you'll have to send me
an email or something.

ES: I'm curious because we're writing again. Like we tried to go and get
away from writing but now we're right back to it again. Like maybe we're
just kind of hard wired to the written word.

CF: I once read somewhere that ever since the invention of language (and
eventually writing) we've essentially been information technology-based
'cyborgs'. I include things like art and culture as types of
technologies - they are tools, extraneous to our natural physiognomy. We
use these artificial technologies (whether it be something as ancient as
language or something as recent as Internet Instant Messaging) to enable
our species to be more powerful than our unaugmented natural bodies and
minds allow us to be. I think that human minds have always adapted to
work with the technologies they regularly interface with, so it's only
natural that we'd find chatting so natural - it's just like writing. And
writing is the bedrock of all other human technologies.



ES: What is it that drew you to an AI interface?

CF: There are many things about AI that interest me. There's the
introspective and existential part: Building (or thinking about
building) an artificial intelligence is a wonderful way of investigating
the nature of human consciousness, who we are and how we know things.
But on another level, AI is, to me, a potentially new art form. So much
of the history of art is about creating artificial persons - with
marble, paint, or with words. I'm not so much interested in AI as a
means for solving problems, playing chess, or being a good secretary. AI
can be so much more than that, it can be fantastic or surreal. Having a
conversation with a single, cleverly wrought artificial personality
could be the equivalent of reading a novel or watching a movie. And
like a movie or a novel, I think that a great AI must be built, in large
part, by an artist with an idiosyncratic vision. I'd like to be like
Stanley Kubrick and make one a cosmically beautiful and timelessly
elegant science fiction movie, only instead of a movie I want to make a
computer game or an artificial personality/celebrity. Technology has to
be a part of it.

ES: So your approach to AI seems like it's tempered with a sense of
literature, which I find interesting.

CF: Ugh, "literature". I mean, I like literature, but I cringe at the
idea that what I am doing is literature. I guess it's kind of *like*
literature. But I wouldn't dare call it literature. I'm not worthy!!

ES: Well, one thing that immediately comes to mind is "Catcher in the
Rye", Which is basically a conversation with Holden Caulfield. You're
asking, "What if you could ask him questions," and I have to say, how do
we know what questions to ask? It would be a very different type of
literature, basically, a literature based on our social skills; social
skills which are said to be going downhill because of things like this
newfangled internet. I love this idea because it is in a sense saying
that socializing, question asking, etc, are forms of interactive
art….Do you think that's legit, or Bullocks?

CF: It's totally legit. One way of looking at a novel is this: the
author creates and defines characters, then forces these characters to
do certain things in order to cause certain situations and certain
events to occur. The events that the author chooses to occur are crafted
based on the author's idea of what the reader will find interesting. A
literature-like bot is really only slightly different from this - the
character's knowledge and conversation skills are designed to lead the
reader (or in this case, the interlocutor) down certain paths and into
certain situations. In other words, I like bots with plots. The bot has
an agenda with you - to make you think about something or to entertain you.

Game designers deal with this every day. Even in a visceral action game
where the plot can be summed up in one word like "survive!" or
"revenge!", the designer has to anticipate the player's actions and
strike a balance between letting the player do anything they want and
letting the player only do the things that actually lead to more
pre-planned situations of carnage. When you've killed all the monsters,
you are free to spend hours mindlessly machine gunning your name into a
brick wall, but the game designer puts in devices that encourage you to
find the doorway to the next level. In the same way, if you ask a Holden
Caulfieldbot about what he ate for dinner last night, it might be more
advantageous for the bot author to divert you back to a topic about
which the bot may have more interesting things to say.



ES: This has me thinking too, of the robots that are out now, for
instant messenger. One of my favorites is VaVaVirgil because it's so
ridiculously flawed, and the icon is a little broken robot, but somehow
it is supposed to prevent kids from smoking. But if you ask it any
questions about smoking it doesn't say anything except "No way! Smoking
is bad for you!" Once I even tricked it into saying that smoking was
cool. I have the conversation saved somewhere. As someone who works with
this technology, is it really possible, given how it works, to create a
seamless interview with a robot, and how soon do you think it will be
until we can do that? I looked at the code for ALICE, [a protocol for
making artificial chat bots] which is basically if/then statements…do
you think we can move beyond this protocol, or do you think we can
really just write so many if/then statements that it comes across as a
real conversation?

CF: Yeah, ALICE is in many ways a hack job. It *can*, however, be more
complex than just if/then statements. For example a conversation can go
down certain pre-planned paths, or the bot can store facts for later
retrieval. But you're right insofar as the theory behind ALICE is that
so much of what we say can be broken down into canned responses and
simple sentence templates.

This conversation we're having, on the other hand, is clearly an
exception. I think we're a long long way off from having a bot that can
do what either you or I am doing right now. I think we need to take baby
steps at this point. Making a bot that is even *interesting* to talk to
would be an accomplishment in my book. I am convinced that AIs will
become a major part of the physical and psychic landscape in which we
live. In 20 years, they will be literally everywhere, so why not start
dealing with the issue now and help influence how they manifest themselves?

I'm not so crazy about the idea of simply creating artificial people, at
least not the part of it that is meant to simulate real people. I want
to make AIs that are more like what we can or might be, that is, barely
recognizable versions of what humans are now living in a kind of strange
amphibious digital/physical worlds, a world where our own minds and the
artificial minds surrounding us are hard to distinguish. Mind you, I'm
not being utopian here, I don't necessarily think that dissolving our
minds into the internet is such a great idea.

ES: It sounds kind of like a nightmare don't you think? The post-human
doesn't really appeal to anyone anymore, does it? It's the new version
of "flying cars by 2001."

CF: I don't like the term "post human" because it implies that we are
now 'pure' humans and that when we interface with computers past a
certain point we will cease to be human. Again, like my "cyborg" example
earlier, we're already there and we just don't notice it. I'm not a
techno-utopian or anything, I just think that there are certain
inevitabilities about how we will increasingly interface with
technology. How technology will evolve - and how *we* will evolve to
work with new technologies.

Ray Kurzweil is famous for using Moore's law to extrapolate exactly how
soon the "flying cars" of machine intelligence will come. But Jaron
Lanier has put out a convincing counter argument in which human
stupidity holds us back from the techno- utopia. How bad design,
political uncertainty, corporate bottom lines, and you know,
'unpredictable events' will change how, when, and if we integrate with
new technologies. Who knows, we could spend 50 years with VHS Artificial
Intelligence instead of Betamax Artificial Intelligence just because of
some stupid marketing decision somewhere.



ES: How soon do you think I could have an Eryk Salvaggio IM bot to take
over for me when I'm at work? So I can have conversations with people
when I'm not even really there. That's something I'm really looking
forward to. That's why I looked at ALICE to begin with, I figured I
could just do a massive list of if/thens and I'd be seamlessly
integrated with my computer; but it wasn't so easy, I gave up. But you
know, I think I would like the idea of a horribly rendered version of my
online personality even better. That's just me, I like technology at
it's clumsiest stages, and as they get slicker and more lifelike I find
they loose their chaos element that can make them so interesting.

CF: I am 100% in agreement with you here. I think that the gaps in a
AI's ability to simulate a real person - that's where the real interest
is. You could easily make an Eryk Salvaggio bot. While it could never
replace the real you it could certainly be an interesting artwork, flaws
and all. I love it when technology fails. It gives you insight into how
the technology works, what it can and cannot do, how the creators chose
to compensate for - or to exaggerate - the flaws. It's certainly fertile
ground for some great jokes.

ES: I know we've talked about this before, about the generation of kids
who loved nintendo games for the immerse world they offered, and the
kids like me that loved getting dust in the heads so that there would be
massive black spaces in the game. Which camp do you think you fall into,
and why?

CF: Oh, both. I don't think they're mutually exclusive. There is not an
artform in the world where the technique of making the artwork isn't
critical to my appreciation of it. I like to be in the world of the
work, and to step out and see how it's made, too.


ES: And you used to be the guy making the video games. You've talked
about subversive stuff you were doing in your video game programming
gigs, anything you're allowed to talk about?

CF: It's not as cloak and dagger as all that. Maybe I exaggerated when I
said "subversive", but in the adolescent-minded world of video games
even such things as an intelligent female character can be considered
radical and challenging. Just before the web boom, I was a video game
designer for several NYC-based game companies. I spent two years working
on an epic/comedic cyberpunk game called "FutureNet". It was never
published - in part, I think, because it was just not commercial enough.
No bloodshed, no cartoon characters. I approached my game design ideas
with almost the same approach I would take to making artworks, that is,
with the idea that my audience would be my intelligent friends and
art-world cultural critics. I thought that I could get paid a salary for
building idiosyncratic artworks. I still haven't learned my lesson,
because I'm still trying to develop an art production 'situation' where
my commercial and artistic enterprises blend together.

ES: So here's something I never understood, and I think most people
outside of the games teams never get, either, is that if you spend two
years on a video game that doesn't get published; what happens to the
video game? Like, does the main code still exist somewhere? Do you have
like CD copies of it or is it just not yours anymore? I was always
curious about "Where Video Games Go When They Die."

CF: Haha, well, I think Futurenet was hurriedly translated and sold to a
German magazine for $8000. It still had a lot of bugs in it, too. I have
a CD copy of the game and I still have some of the source files, but
otherwise it's in limbo. It's heartbreaking that it didn't get published
- it was a kick ass game.



ES: So you told me that the Rhizome Grant is one of the first times you
got paid for a pure work of art. How does it feel compared to the Video
Game gigs?

CF: Well, even the Rhizome commission has a wee bit of a "work for hire"
feel to it. I mean, it's not like Rhizome would be all that happy with
my final product if it did nothing but hurl insults at all of the
artbase artists or at Rhizome.

ES: Oh my god, that would be great. If I was working at Rhizome I'd be
totally into that.

CF: Not that I feel constrained in the least, though. After so many
years of doing work for clients where my creative range is severely
limited, it's great to work for a project where I can pretty much do
anything I want. I could probably even make the insult-hurling bot, but
I don't think I will. I see the work as a kind of collaboration with
Rhizome, a collaboration where they've given me more creative freedom
than anyone who's ever given me money has allowed me. It's great. I wish
I could get paid for all my art ideas - I've got about a thousand of
them ready for production, and thousands more of them germinating in my
head.


ES: On your site, you say the distinction between art and science is
really kind of a joke. I guess I am wondering if you could elaborate on
that, and how your work addresses that.

CF: One one hand I condemn the art/science schism, but on the other hand
I maintain two sites (askrom.com and graphpaper.com) that perpetuate
that very dichotomy. Http://askrom.com is my "technology laboratory"
site, and http://graphpaper.com is where I develop visual, conceptual,
and personal explorations. I'm pretty up front about how askrom.com is
the "science site", and graphpaper.com is the "art site". It's sometimes
convenient to have 'compartments' in which to drop my work. I'm
interested in art and science for two reasons. First, because I believe
that art is something that is worth talking about in concrete terms,
even in scientific terms. And second, because I love science and
technology and I want to make artworks that use both. Oh, and third, I
think the "gaps" between art and science are incredibly fertile ground
for bitter critique and hilarious satire. A lot of my work is about this
very gap: my absurd inability to use science to explain social,
existential, or aesthetic phenomenon, my own inability to even
understand certain sciences at all.

A really ambitious art project would be to try to create artworks which
attempt to illustrate the greatest and most challenging scientific
concepts of our time, like string theory or cognitive modeling. What
better way to celebrate human creativity, artistic and scientific, than
to try to use art to help people bring some understanding of big science
into their lives?


ES: I notice that you tend to present a lot of sketches of potential art
projects.

CF: Oh, god yes. I'm about to post a couple hundred new sketchbook
pages, each of which is a potential project. Making technology-based art
is doubly frustrating. While some net.artists are content with the DIY
aesthetic, I think my professional background in computer game design
and information architecture has allowed me to have ambitious art
project ideas. I just don't even get ideas that wouldn't require at
least a few weeks of dedicated time to build. A good deal of my ideas
would require five- or six- figure budgets to realize. So instead of
making a half-baked version, I just make a nice drawing of the work and
move on to the next idea.

ES: Is that satisfying for you?

CF: Well, I love drawing. Secretly I have a dream that I can have a
career that structurally resembles the career of Claes Oldenburg. He
made tons and tons of drawings, and occasionally got an opportunity to
build something big, to see a drawing realized. But I always got the
feeling that he was happy with just the drawings. He draws a giant
faucet and he's happy. His "unrealized" drawings don't feel incomplete
to me. I really like this approach. This is why my sketchbook isn't just
a place to jot things down. I spend a lot of time with fine point
colored pens, designing and crafting the visual language around an idea
that might have worked as just a few scribbled words. I polish my
doodles because I am painfully aware that the polished doodle may be the
end of the line for each idea, that it most likely will never make it
all the way to completion.


ES: Have you ever come across a site that uses a portion of your code,
and felt a sense of ownership/involvement, when you really never had any
idea what was being done? Or does that just piss you off? There's an
interview with Aphex Twin where he basically complains that all his most
interesting, experimental ideas he's not releasing, because he's too
afraid of getting ripped off. He'd rather just see what people came up
with on their own.

CF: I'm quite reckless in this regard. People are welcome to steal my
code. Half of my code is stolen from somewhere else anyway, although I
generally give credit. I don't consider my technical skills to be
anything special. There are literally millions of great programmers out
there today who are better than me, and half of them are probably still
in high school. No, I don't worry about people using my code - in fact,
I'm a bit disappointed in how rarely my little free FLA's are
downloaded!

What sometimes worries me is the fact that I have my sketchbook online.
Like I said, my sketchbooks contain the earliest drafts of almost every
idea I ever have. In a painful episode for me, I recently "caught"
somebody who I am 99% sure made it through their junior year of art
school stealing ideas from my sketchbook at graphpaper.com. But I say
fuck it: for each idea I put online there's another, even better one
right behind it. And I want people to see the ideas gushing out.

By putting my sketches and my source code online and sharing it with the
world, I feel like - well, no, I really AM - part of a greater online
community. I don't understand how a technology-based artist can live
with themselves without having a kick-ass web site these days, sharing
their ideas and linking to other online friends. I don't think I would
have ever met a good number of my current friends if I didn't have a
cool web site.


URLs:

Christopher Fahey
http://www.askROM.com
http://www.graphpaper.com

Rhizomebot Proposal
http://www.graphpaper.com/rhizomebot/

A.L.I.C.E. AI Foundation
http://www.alicebot.org




++++