Information Technology Reference
In-Depth Information
to themselves, and by virtue of your similarity, probably useful to you too (although
by no means definitely). Adaptive imitation of behaviour is a particularly human ca-
pability, and a challenging cognitive task (Conte and Paolucci
2001
). The imitation
of successful behaviours allows human social systems to be cumulatively creative,
amassing knowledge and growing in complexity (Tomasello
1999
). Secondly, as
discussed in Sect.
14.3.2.2
, social structures bind individuals together into mutually
adaptive behaviours: John Harrison did not build clocks so that he himself could
better tell the time at sea.
How this common or mutually adaptive value works in the arts, however, is less
clear, since the value of any individual behaviour is determined not with respect to
a static physical environment but a dynamic social one (Csikszentmihalyi
1996
).
Whereas the value of washing food does not change the more individuals do it,
the value of an artistic behaviour can change radically as it shifts from a niche be-
haviour to a mainstream one. In this way, copying successful behaviour does not
necessarily lead to an accumulation of increasingly successful behaviour, as in the
accumulation of scientific knowledge, but can also lead to turbulence: unstable so-
cial dynamics predicated on feedback. The value of artworks to individuals is highly
context-specific and suggests this kind of dynamic. Thus it seems more appropriate
to look to the second way in which adaptively creative systems can be of use to
others, by being locked into mutually beneficial goals through social structures, but
also to recognise that copying successful styles is an essential part of this process.
The arts appear to involve
ad hoc
groupings of individuals who share common goals,
into which adaptively creative arts-based computational systems could become inte-
grated and be of benefit to individual humans. This points to the idea that achieving
success in arts-based computational creativity is as much a matter of establishing
appropriate individual and social creative contexts, practices and interfaces as it is
of designing intelligent systems.
The
Drawbots project
investigated the idea of producing physically embodied
autonomous robot artists which could honestly be described as the authors of their
own work, rather than as proxies for a human's creativity (Bird and Stokes
2006
).
This would have overcome the limitations to the agency of the software in examples
such as Harold Cohen's
AARON
(McCorduck
1990
), where Cohen is clearly the
master of the creative process, and AARON the servant. The project illustrated the
fundamental conundrum of attempting to embed an artificial system into an artistic
context without proper channels through which value can be managed. In fact, the
drawings produced by the Drawbot were no more independent of their makers than
AARON's, and arguably had less of a value connection to the outside world than
AARON did, even though AARON's connection was heavily mediated by Cohen,
its proverbial puppeteer. The Drawbots possessed independence in a different sense,
in so far as they were embedded in their own artificial system of value. Thus each in-
dividual Drawbot was individually adapted (the product of an evolutionary process)
but not adaptively creative, and the entire system was generatively creative (able to
lead to new patterns and behaviours) but also not adaptively creative, and thus not
creative in the sense of a human artist.
