<nettime> The Human and the Machine: interview with Nick Baginski

[Josephine Bosma <jesis@xs4all.nl>]

The 28th of June '97 there was a little gathering in V2 Rotterdam as
part of "Machine Aesthetics". It was an artists workshop, next to an
exhibition, some presentations and a seminar, which took place the next
days. One of the artists that presented a work was Nick Baginski, who
has been working with artificial intelligence and neural networks for
many years now. His work here was a guitar playing robot, that gave a
long hypnotizing solo.
I met Nick for the first time in 1994, when he had an installation in
the old V2 building that was called "Survival in Bosnia".
This was a projection of a computerscreen on a wall, on which the map
of Bosnia was occupied by three groups and colours of dots, artificial
life forms, that fed, lived and killed/ate eachother, a blue group of
dots available for the audience to place in between the competing groups
at will, but without ability to stop the fighting completely ever.

We talk about creativity and artificial life, the human response to
machines that have certain human 'qualities' and the impossibility of
ethics.



JB: I haven't talked to you since your survival in Bosnia project.
Can you tell me what you have done between then and now? That was
also an artificial art work, right?

Nick Baginski: That was an artificial life piece based on genetic
algorythms. I've done many many things since then, I couldn't recall
them all now. I've been mainly working on the fusion of artificial
intelligence and music. That was a very important topic in the past
years: how to build artificial mechanisms that can generate music,
that can learn about the nature of music, the fysics of music and
then apply this learning to playing an instrument and generating
music.

JB: Does that mean also the emotional parts of music?

Nick Baginski: I choose music because music is the artform that goes
straight into your heart. Thats the special thing about music. Its
been less catagorised and less thought over on reception then any
other artform. Thats why it's ideal for combination with so-called
artificial intelligence, the gap is the biggest. The contrast between
what it does and who does it is the biggest. That's what I think makes
an interesting statement or something worth thinking about.

JB: What is that statement?

Nick Baginski: I find it rather interesting to find a (compared to
biology) rather simple mechanism, that has the ability to generate a
musical system of its own. This musical system is not so far from what
we think music is, because in both cases its build on the nature of
the fysics of music and the law of how sound progresses in air. Working
on this told me that our musical achievements, our knowledge about
music are very much dominated by the fysics behind it. Even a rather
simple mechanism can do not necesarely similar, but close related
things without having the abstract symbols, just by listening and
playing and trying out. It puts quite some new light, I think, on our
creative achievements.

JB: The media dictate the outcome?

Nick Baginski: Not necesarely the outcome, but the system that's being
generated around it. This thing can play a blues scheme in E-major
without knowing about these things. It plays E-major because E-major
is very logical to a guitar and to a guitar tuned in E that has the
ability of playing slide. It actually more or less has to play a blues
in E, because it IS a guitar: it has an open tuning, it has a specific
sound, it can't do anything else. Maybe there you can draw the line
between such a, what I call, creative mechanism and the human being.
A human being can, allthough he is holding a guitar, still play
something completely different on purpose. But naturally a guitar,
depending on what kind of guitar it is, plays either blues or flamenco
or whatever we know is logical for a guitar. This is demonstrated here
by this robot.

JB: You have also build other machines. Did you see the same mechanisms
there?

Nick Baginski: There the intentions were different. For example there
is a group of little robots I built, that are partly disabled. They are
autonomous battery operated little creatures. I call them creatures,
I may (laughs). They can't do anything really at all. They are not
succeeeding in walking or doing actually something. They are meant not
to be succesful. They are meant to provoke some emotional reaction
like pity, feeling sorry for them etcetera and it was very interesting
to watch how people would respond to these very simple patterns of this
fysical being that tries to walk, but does not succeed in walking. It
tries really hard.
People always try to help them. The machine does not care if its not
succeeding or if some human is helping it. The human actually gets
something out of helping this thing. It is interesting to watch these
mechanisms. I made these robots in the streets and people give them
money. They don't give them money to help the robot, that does not make
any sense...they give them money because it makes them feel good.

I really like to uncover these mechanisms: how we work, how we operate.
I use artificial intelligence for this. A robot is a very good medium,
because its neutral.

JB: so your machines tell us more about our own emotional mindset then
about the aesthetic of the machine itself?

Nick Baginski: As always, this is working on many levels. I do this
work because I learn something about my musicality and musicality in
general, but I also do it because I learn something from the audience's
response. Its very important how the audience actually responds to this
piece. That is being reflected in the title. The piece is called the
three sirens, as originally there were three robots, here we only have
one right now. They have a way of continuous playing, because they don't
know any song structures like a song that has a begin, middlepart and
hookline and things like this. It keeps playing for hours and days.
Sometimes people get caught in exhibitions. They stay here for an hour
or half an hour and listen to this continuous flow of music. Its really
hard to get away from it. Its captivating after a while. To watch this
is important to me, but also to learn about myself and social mechanisms.

The aesthetics are not machinic, the aesthetics are dominated by the
purpose. By the purpose to build something that performs in front of an
audience. Its musician aesthetics and not machine aesthetics that I am
in this case interested in. Earlier, with the little quasies, there I am
interested in criple aesthetics. Its not the machine itself, thats not so
important for me.

JB: In another work you sampled one face from a whole audience of faces.
Also you did something with their voice. Can you tell us what this was
and why you made it?

Nick Baginski: I just made this observation that no matter how
complicated a piece or artwork is set up, the technical structure or
whatever, even the decoration in a shop window, as soon as there is
an image of a visitor from the audience appearing on the screen,
everything else is not important anymore. People get caught immediately
by their own image. This narcistic mechanism I find very interesting.
In this case I on purpose give the audience a choice: they can have a
look at the art (there is one projection of the artwork) and there is
another image, of themselves.
This machine again is a quite complicated neural network mechanism
that filters out, finds faces in a video image. It is then played
through the video projection.

JB: I understood that this machine made also a kind of assemblage of
all the faces it saw. It makes one face out of hundreds of faces.

Nick Baginski: That is a side effect of the learning strategy, to find
sort of general faces, role models of faces. It does this by combining
or clasifying everything it sees, namely faces, and then combining the
ones that have similarities. By doing this it generates quite strange
images of noses and mouths of different people being joined together
into one image.

JB: What did you get out of this project for yourself?

Nick Baginski: It's almost learning mechanisms I study. In this case I
was interested in building something that could recognise significant
properties of faces, which could be anything. The recognition of faces
by human beings is a very special achievement. We have a very broad
area in our brains that is just specialised in recognising peoples
faces. Even if there is a person we know out of our private circle
and we are shown an image of them with 80 % of the image blacked out
and we just see one eye, we still immediately recognise that person.
That is a very high cognitive feature we have. That is why recognition
of faces attracted my attention. It's one of the most important
mechanisms in our daily life. Communication is a vitally important
thing to us.

JB: When I was interviewing you years ago, with "Survival in Bosnia", we
were talking about if it were possible for artificial creatures to live
on the net. Finding their own ways about.. That of course was a very
romantic notion in some way...

Nick Baginski: It now is reality. The definition still is the problem
of course. I have learned to point out one important thing that people
tend to forget. Artificial life, and the same implies to artificial
intelligence, is something completely different from real life and real
intelligence. That you have to remember. When you look at autonomous
agents roaming through the net and collecting data or lets say simple
search engines, are not necesarely life forms, but they have a very
high degree of autonomy and sort of life in the net.

JB: What you do now, would you call these autonomous life forms?

Nick Baginski: No, I wouldn't. Again, the definition of life is always
problematic, but reproduction and self support are two very important
aspects of life. This installation does neither reproduce, nor does it
support itself, it needs my help desperately.

JB: You say that Artificial Life and what we call life cannot be compared
really. Does that mean that the life forms that you see in networks and
so on cannot be compared to early, lets say natural life forms like
amoebe?

Nick Baginski: I don't dare doing it because the medium is so completely
different. You cannot compare a computer memory to the prehistoric
situation on earth. There is just no way of finding a scale to compare
these two things. Similar is the outcome really hard to compare. The
first amino acid structures compared to lets say a virus or a search
engine robot on the net, I see no measurement, any dimension where these
two things can be compared.

JB: So a search robot can be called artificial life...

Nick Baginski: I don't know. I believe there are much more clever
autonomous agents on the net then the ones that are collecting the
information for search engines. About virusses: it is not even defined
in real life, in biology, whether a virus is alive or not. It does
reproduce somehow, it does not have a meta exchange of its own..so, is
it alive? Its a question of definition. It is very close to what we
call life, at least very close to artificial life, thats for sure.

JB: Why did you object to the name 'Machine Aesthetics', the name of
this exhibition?

Nick Baginski: Its a too universal concept. It can be anything. Thats
why I am having difficulties with this term. It can be something like
the post industrial aesthetics of the italian futurists. That is machine
aesthetics at its best, at its most original form. Or the drawings of
Leonardo DaVinci are machine aesthetics. The design of a search engine
or an artificial intelligence program is also machine aesthetics, and
I don't see any common criteria to compare their aesthetic concepts.
The term machine aesthetics is just to wide open to define anything.

JB: How would you call it? This exhibition here?

Nick Baginski: The topics being discussed here and the pieces shown
are very diverse. Maybe we could talk about the aesthetics or
beauty of information processing. This might be a term that all the
pieces shown here apply to. I think information processing is a very
central thing here. Like for instance TV poetry next door, which extracts
random poetry out of random tv images, (by Gebhard Sengmueller. JB).
It scans written texts out of tv images, running programs, and then has
an artificial voice speaking these lyrics, these poems to us. That is
a machine, a mechanism working and generating somehow an aesthetic
output.
My piece does a lot of information processing, by hearing itself and
generating new music out of what it hears. The exiting thing is not the
truely mechanical visible thing, but the beauty of the information
processing.

JB: With artificial life, could we go to something like machine ethics?
Could we discuss machine ethics?

Nick Baginski: Maybe its good to make one point in saying that all
the machinery we have, we build, I would call primarily prosthesis.
The first weapons were invented to prolong your arm, strengthen your
power in fighting against your enemies or whatever. The knife was
invented to increase your level of capabilities. All these things are
prosthesis. They extend the capabilities of your body. If you look at
the world around us, bearing this in mind, its very interesting to see
the car as part of your body, which it actually is. You don't become
part of the car, its the other way around: the car becomes part of
you. It is replacing your legs. Bearing this in mind, looking at the
aesthetics of it, it is very obvious that everybody decides whether
the new legs you have sort of fit the aesthetic concept of yourself.
So its mostly your personal aesthetic concept that dominates the
selection of tools or devices you add to your body.

JB: With ethics its the same?

Nick Baginski: Its the same. If you want yourself to kill: if not you
would not add a gun to your body. So the ethical question is solved
from the very beginning, whether your relationship towards a weapon is
positive or negative. Do you want to have a gun as part of your body or
not, that's pretty simple to answer for everyone.

JB: There is a little contradiction in what you say, because earlier on
you said the media dictate the outcome. It's of course so that if you
add a knife to your body, or a gun or a car, they all have very specific
features and specific results. Therefore maybe it is possible to talk
about the other way around, not seeing them as prosthesis, but also as
dictating something upon the human that uses it.

Nick Baginski: We are forced nowadays to own a car. You don't have to,
but there are many good reasons to do so. Still you can say: no, I
don't want to have a car, I don't want to deal with this. My legs are
fine, I am satisfied with this. I don't need to add wheels to my body.
Of course it gets more complicated when we start talking about
collective responsibilities like a country has, or a group of people
has, because then you can't say: I don't want this, because you don't
have the individual body or language anymore. That's a problem that not
only applies to machines, thats a problem that applies to every decision
you have to make. Its no new quality in an ethic discussion.

JB: What does add a quality to an ethic discussion is: do you make a
machine yes or no? And: how do you make it? Thats a question I haven't
heard here, that's a question that was not discussed.

Nick Baginski: Because that is the decision of each individual in the
end. Do you want a car, do you want a knife or don't you? There's no
principally new question in it in this case.

JB: But you have build a wonderful machine here. It is truely hypnotising
as you said. When I first saw it I said to the person I was talking to:
"I can easily imagine a crowd of ten to fifteen thousand people cheering
in front of it". It's very Hendrix at his most experimental.

Nick Baginski: Only the performance so far is a little bit poor. Its ok
for a guitar, but if you look at the needs of a band, you need a real
frontman, one that really does the emotional work on the audience. Thats
a very complicated concept. I have been thinking about singers and
building a person that does this job, of capturing the attention of
millions, so to speak, by his mere appearance. I haven't solved this yet.

Nick Baginski:
http://www.provi.de/~nab

V2:
http://www.v2.nl


*
---
#  distributed via nettime-l : no commercial use without permission
#  <nettime> is a closed moderated mailinglist for net criticism,
#  collaborative text filtering and cultural politics of the nets
#  more info: majordomo@icf.de and "info nettime" in the msg body
#  URL: http://www.desk.nl/~nettime/  contact: nettime-owner@icf.de