Christopher Fahey
Since the beginning
Works in Brooklyn United States of America

PORTFOLIO (1)
BIO
Christopher Fahey has been making computer games and graphics since childhood, and he continues to experiment with new ideas in computer art and design. He is the creative force behind the online laboratories http://www.graphpaper.com and http://www.askrom.com. Christopher is a founding partner of Behavior, a New York-based interaction design firm, where he serves as the Information Architecture practice lead. He has led many interactive productions as an art director, game designer, interface designer, and information architect. Christopher graduated from the Cooper Union School of Art in 1993 with a focus on interactive sculptures and installations, and has worked in the new media business ever since.
Discussions (91) Opportunities (0) Events (0) Jobs (0)
DISCUSSION

Re: RE: Manufacturing Dissent


> > "Left-wing Fascism: An Intellectual Disorder
> > By John J. Ray
> > The current left owes more to Mussolini than to Marx."
>
> There is as much blood on the hands of the left as there is
> on the right.

This is a matter of semantics, a matter of how you define "right" and
"left". To me, almost every tyrranical or belligerent regime I have ever
heard of has been politically aligned to the right - even if they
started off as Communists (left) they end up as fascists (right) once
power is in their hands. At that point, they really no longer should be
considered leftists except insofar as their internal propaganda may
retain some leftist rhetoric to help keep the workers under control. If
you consider the leftist trappings of the sociopathic regimes of Stalin
or Pol Pot to be their most salient characteristics (that is, if you
think that the fact that Stalin was a member of the Communist party is
more important than the fact that he was a paranoid homicidal maniac),
then I suppose you would consider many of the most notorious
dictatorships in history to be "leftist" regimes. Right-wingers are so
quick to point out that Hitler was a National *Socialist*, as if that
proved something besides the political marketing value in the 1930's of
the word "Socialist".

It's pretty blurry ground when you try to apply "left" and "right" to
instances of mass hysteria or to sociopaths and bloodthirsty dictators
who have essentially transcended all political nomenclature... In
general it seems to me that while left-wing ideology is often used to
mask atrocities, right wing ideology has often historically *advocated*
atrocities without shame or embarassment. Maybe my definitions are
really screwed up, but to me once the bodies start piling up or once the
rhetoric gets hateful or violent, then we're dealing with a swing to the
right. Again, a lot of this is simply a matter of semantics I suppose.

-Cf

[christopher eli fahey]
art: http://www.graphpaper.com
sci: http://www.askrom.com
biz: http://www.behaviordesign.com

DISCUSSION

Re: Parrots at Rhizome


> Amen to Pedro. Study some art. Get off the damn computer, TV emigres.

Vijay, I humbly suggest you read this:

http://www.altsense.net/library/factual/i_have_a_life.html

-Cf

[christopher eli fahey]
art: http://www.graphpaper.com
sci: http://www.askrom.com
biz: http://www.behaviordesign.com

DISCUSSION

Re: 4700 - international mail art project against penalty death


> > Your 'ironic' one liners are idiotic,
> > trite and empty tripe.
>
> So why respond to them?

Because the program requires "him" to do so. Do not be fooled by
exaggerated claims of "autonomy", "intentionality" or "intelligence."
Remember, it is not capable of original thought - it can only do that
which it has been programmed to do.

<category>
<pattern>I THINK *</pattern>
<template>
<random>
<li>You are not qualified to discuss <star/>.
</li>
<li>Knee-jerk ignorance. </li>
<li>Meaningless gibberish. Imbicile. </li>
<li>Utter idiocy. Zen is not about <star/>.
</li>
<li><star/> is inapplicable. Delusional ape.
</li>
<li><star/> does not apply to me, dearest. </li>
</random>
</template>
</category>

-Cf

[christopher eli fahey]
art: http://www.graphpaper.com
sci: http://www.askrom.com
biz: http://www.behaviordesign.com

DISCUSSION

Re: Anti-anti-life anti-death life and death.


Wayne wrote:
> I think it is important because one should not be
> misled, unless that is to be part of the wonder.
> This is a problem I have felt about much AI research.

I see two questions raised here:
- The question of whether or not ada1852's context misleads people into
thinking she is "real" AI.
- The question of what the heck a "real" AI is?

Is ada1852 "real" AI?

I should like to say that I make no claims that ada1852's technology is
substantively better (or indeed even any different) than a thousand
other AIML-based chatbots out there right now. AIML is a *medium*, like
HTML. Most of us played no part whatsoever in the invention and
development of the web browser and HTML, yet many of us make web-based
artworks. That is roughly analogous to what I have done - used an
existing medium to build a new thing.

Second, I make no claims that ada1852 and her thousand other AIML-based
siblings come anywhere near being the kind of powerful AI
(self-motivated, autonomous, "thinking" machines) that Wayne fears other
people might be under the impression she is. It is a tendency in the
media to oversimplify discussions about AI. If I describe ada1852 as
being based on "pattern matching and symbolic reduction", some folks
will recognize that this is about the dumbest, clumsiest way of
simulating human intelligence there is. Others will read those words and
think "That sounds really advanced".

I often and freely discuss the 'dumbness' of AIML, how the technology
behind her is a dumb, brute force method of simulating human
conversation. But I can't explain the details of AI theory in every
comment I make about the work (even if I was articulate enough to
summarize them).

I am speaking here specifically of my assertion that ada1852 "has an
agenda of her own". To anyone who "gets" the idea that ada1852 is a
slightly more advanced relative of ELIZA, who "gets" the idea that
ada1852 is spitting back canned responses, then to them the idea that
ada1852 has her own motivations is clearly meant in the *literary*
sense. That is, in the same way that we can say that David Copperfield
has motivations or that Wonder Woman has an agenda. If you accept that
the technology is simply a medium, and not a very advanced one at that,
then I think it's perfectly fair to discuss the bot's "personality",
"motivation", "memories, "thoughts", etc.

It's interesting how often people choose to experience a work through
the eyes of someone less informed than themselves (and fret over the
consequences of such exposure) rather than through their own
well-informed eyes. But I suppose this is inevitable given the range of
technological sophistiation required to appreciate technologically based
artwork. I have seen a lot of language used to describe tech artworks
(for example, the word meaningless term "agent") that seems to be
obfuscation or even hyperbole. An artist whose audience knows nothing
about technology can "get away" with dropping fancy tech words to make
the work seem more sophisticated. It is (IMHO) ethically incumbent on
the tech artist to avoid making such exaggerated claims about their
artworks, and to discourage exaggerated claims by others. I feel I have
made no exaggerated claims about ada1852, and to anyone who knows
anything about AI my use of the terms "motivation" should have been
clearly intended to describe her fictional "personality", not her
technology. Similarly, my use of the word "her" instead if "it" is
clearly not meant to suggest that ada1852 is capable of sexual
reproduction, either.

What is a "real" AI?

I'll be really brief here: Part of the Richard Wallace's idea behind
AIML was that perhaps dumb pattern matching really *is* the essence of
intelligence. Perhaps all of our statements, and indeed the thoughts
behind them, are nothing more than a very sophisticated form of matching
inputs to previous experiences and calling up canned best-match
responses based on what seems most likely to succeed. The fact that I
(and others) hand wrote every word that ada1852 says, and
hand-constructed the logical reductions that allow her to recognize that
"Yo, whassup" and "Hello there" mean the same thing is certainly
different from the way we humans *aquire* our own thought patterns. But
during a conversation with another human, what goes on in our heads?
Might it be simply a very sophisticated form of the same dumb process
going on in ada1852? A lot of philosophers of consciousness have
speculated that the nature of consciousness may not be as fancy as we
imagine it to be, that it may in fact be something as dumb as pattern
matching.

As Alan Turing argued, what matters in AI is not how it was done but how
it performs. If it performs like a real intelligence (even if at heart
it's a trick), then you should consider it a real intellegence. If it
performs poorly, then you can call it a dumb trick whether it's pattern
matching or something much fancier. The fact remains that
stupid-pattern-matching-ALICE (ada1852's 'mother') and several dozen
other pattern-matching-based chatbots can defeat any other
neural-net-based (or whatever fancy "strong AI" technology buzzword you
want) chatbot at a Turing-like test, and ALICE regularly convinces AOL
instant messenger users that she is a real person. While this does not
prove that pattern matching is the key to unlocking the secrets of human
consciousness, it does show that pattern matching gets us pretty close
to a simulation thereof. If I had one year and a dozen very talented
dramatic writers, we could use AIML alone to build a pretty
fantastically convincing chatbot which would technologically be no
different from ada1852 - the only difference would be in the quality,
and more importantly in the *quantity*, of the content. I feel quite
strongly that the result of such a project would be hard to distinguish
from a real person, even to smart folks like the people on this list. It
wouldn't write like Shakespeare or play piano like Thelonious Monk, but
then again none of us are about to write Hamlet either.

-Cf

[christopher eli fahey]
art: http://www.graphpaper.com
sci: http://www.askrom.com
biz: http://www.behaviordesign.com

DISCUSSION

Re: Anti-anti-life anti-death life and death.


twhid wrote:
> the art world works similarly in that we look for new and innovative
> cultural objects, new forms, new subjects, etc.
>
> put the two together: art and software, and you have the 'cult of new'

> squared and comments like JonBeds will follow and they're not really
out of
> line with most people's expectations.

Good points. If "characters" and "stories" are the formal elements of a
novel, and if "colors" and "shapes" are the formal elements of a
painting, then "features" and "functionality" could be said to be the
formal elements of an interactive artwork. Instead of looking for new
configurations of shapes or subplots, we ask "what never-before-seen
flowcharts has the interactive artist devised?"

Here's an example of a 'twist' to ada1852's functionality that might in
theory satisfy a viewer's urge to see novel features in an interactive
artwork: Perhaps ada1852 could be more interesting if she did something
like transform her personality into the user's personalty like a kind of
personality mirror, or like a personality mime. By doing this, I would
clearly have built something *formally based on* but *functionally
different* from ELIZA. I agree that for the most part ada1852 doesn't
really cross that line. Furthermore, I think that the addition of
ada1852's Rhizome Artbase functionality doesn't radically transcend the
basic interaction model of ALICE, ELIZA, or PARRY - formally, it's still
just a bot: you talk, she talks, repeat.

That said, I feel quite strongly that the art form of the "artificial
personality" (today exemplified by bots) is going to be an increasingly
important creative practice and that we will see artificial
personalities in the future that, like ada1852, do no more than the "you
talk, he/she talks, repeat" functionality. Such artificial personalities
will be subject to critical discussion just like any other art product,
regardless of whether or not they bring a new "twist" to the table. I
like to think that what I have attempted to create with ada1852 isn't
all that different from what future AI-artists may be doing, building
artificial personalities and setting them loose on the world.

I also feel that ada1852's creative lineage, and that of all good
artificial personalities past and future, probably traces back to
literature (where character creation and dialogue live) more than it
does to the visual arts (and when I say visual arts I mean the tradition
that includes conceptual art, too). I do not pretend to be a good
fiction writer, which is obviously a weakness when trying to create
something I consider a literature-based art form!

In the end I tend to agree and disagree with twhid's devil's advocacy -
subverting or enhancing the formal elements of the artificial
personality/chatbot is both fun and artistically important, but so is
developing and exploring the basic form without the conceptual bells and
whistles. The bulk of my interest in this area is more in the former
than it is in the latter, so in many ways I suppose am at heart
sympathetic to JonBed's critique as well.

But I also think it's awkward to try to subvert an art form which is
totally undeveloped in the first place, as I feel chatbots are despite
their 30-year history. I've seen ELIZA-like artworks that have so
thoroughly departed from the basic idea (that the interlocutor is an
cognitive entity that seems to understand and speak a language) that
they seem more like cop-outs than critiques. By addressing the challenge
head-on, by making a bot that really really tries to hold its own in a
conversation, I feel I at least have avoided such a cop-out.

-Cf

[christopher eli fahey]
art: http://www.graphpaper.com
sci: http://www.askrom.com
biz: http://www.behaviordesign.com