[Synapse elist] Wrapping up June's discussion

[Daniel Bisig]

Hi all,

I apologize for my extended stretch of silence (too many paper  
deadlines had been looming over my head throughout this month). I  
don't think I can and want to wrap up the discussion. One guest  
researcher at our lab once mentioned that he considers AI to still be  
in a stage the resembles alchemistic inquiry much more than a  
established scientific field that is well structured and is founded on  
a canonical body of knowledge. So the only form of conclusion I could  
possibly draw is that AI is a fascinating field because of this very  
weakness of immaturity. And it is this openness and explorativeness  
that renders this field particularly attractive for an art-science  
discourse.

So instead of coming up with some possibly arbitrary conclusions, I  
would like to offer a few thoughts on the most recent discussion going  
on here on the list. This discussion concerns the relationship between  
an artifact's origin and it's capability to create things. Concerning  
biological organisms, there is this nature-nurture debate that is  
going on since pretty much forever and that is not only of high  
scientific interest but has large implications on the very foundations  
of human society. With regard to AI system (such as robots) on the  
other hand, I don't think there is much room for such a debate at the  
moment. The behavior and capabilities of almost all robots I'm aware  
of is totally dominated by their origin and very little have been  
acquired by the artifacts themselves. Ironically, artificial systems  
fall particularly short in comparison with natural systems in that  
they are almost exclusively defined by this debate's nature aspect. I  
therefore think that before we can hope to create a robot that is  
creative (bet it painting, problem solving, communication or whatever)  
we have to come up with concepts and designs that endow robots with  
the capability for open ended adaptation. This means that robots not  
only needs to possess much higher levels of plasticity in its  
morphology and control than nowadays systems do, but much more  
importantly, a robot need to be able to evaluate what it is doing and  
learning in its own terms. It needs to have its own value system,  
drives and motivations. It needs to be able to develop its own  
criteria what is relevant to it and both recognize and appreciate  
novelty based on its own understanding of relevance. I'm strongly  
convinced that there is no shortcut in creating creative robots that  
are anything more than just fancy tools for artists who try to  
identify and automate their own artistic activities. It will then be  
interesting to see if truly creative robots create things that can be  
appreciated by a human audience at all.

Best regards

      Daniel

-------------------------------------------

Dr. Daniel Bisig

Artificial Intelligence Laboratory
University of Zurich
Andreasstr. 15
CH-8050 Zurich
Switzerland
T: +41 44 635 45 77
F: +41 44 635 45 07
E: dbisig at ifi.uzh.ch
W: bitingbit.org

Institute for Computer Music and Sound Technology
University of the Arts Zurich
Baslerstrasse 30
CH-8048 Zürich
Switzerland
T: +41 43 446 55 18
F: +41 44 578 78 10
E: daniel.bisig at zhdk.ch
W: bitingbit.org

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.synapse.net.au/pipermail/elist/attachments/20080630/be2b2e2c/attachment-0001.html