[Synapse elist] humanlike

[Paul Brown]

An interesting observation here from the psychology of robotics.

Researchers have discovered that robots that are too human-like  
generate unease and fear in humans (too much of a threat maybe - plus  
the "terminator" myths).  Robots that are not designed to simulate  
human features/build are much more readily accepted.

Also worth looking at the work of Owen Holland in the UK.  See http:// 
cswww.essex.ac.uk/staff/owen/ and read down the page for his comments  
on anthropomimeticism.  He uses an interactionist argument based on  
embodiment that in order to develop human-like intelligence you need  
to have human-like functionality and behaviour.

Paul

On 2 Apr 2008, at 19:13, Leonel Moura wrote:
>
>
> Considering Kirsty's question: how do we want our robots to look  
> and act?
>
>
>
>
> I see robots as a non-biological life form. Hence, the more
> independent they are from us the better.
>
> In the same way that bio-inspired art shouldn't be reduced to
> vitalism, robotics, applied to art or any other activity, cannot be
> seen just in the context of anthropomorphism. Autonomy, evolution,
> reproduction cannot be achieved by machines (probably biomachines) if
> we follow a top-down process stemmed from human features. Like in
> nature we must look for the edge of chaos.
>
> The other point is that natural beings are not just the product of
> their genes and body, but also of their interaction with the
> environment. Building robots as an assemblage of machinery and
> software to mimic human behavior is less interesting than to generate
> stochastically the basic conditions for a non-biological entity to
> survive by its own means. Situatedness, one of Brooks concepts, is
> for example essential when considering true autonomous robots. We
> cannot do that with control.
>
> In short, to build a bacterialike robot seems to me much more
> exciting than to build a humanlike robot.
>
> And as an artist my question is less how it looks like, but will it  
> work?
>
> _______________________________________________
> elist mailing list
> elist at synapse.net.au
> http://lists.synapse.net.au/mailman/listinfo/elist

====
Paul Brown - based in OZ Dec 07 - Apr 08
mailto:paul at paul-brown.com == http://www.paul-brown.com
OZ Landline +61 (0)7 5443 3491 == USA fax +1 309 216 9900
OZ Mobile +61 (0)419 72 74 85 == Skype paul-g-brown
====
Visiting Professor - Sussex University
http://www.cogs.susx.ac.uk/ccnr/research/creativity.html
====