Schizophrenia and narrative in artificial agents - Narrative Intelligence - page 277

Conclusion
In this paper, I have argued that there is a fundamental lack in autonomous
agents' behavior, which reduces their apparent intentionality. By being con-
structed in a fragmented manner, agents suffer a kind of schizophrenia, a
schizophrenia which can be addressed, in analogy to anti-psychiatry, by mak-
ing agents narratively understandable. In order to do this, I have built an
agent architecture which combines (1) redefinition of behaviors as signifiers
and their reorganization in terms of audience interpretation, (2) the use of
transitions to explain agent motivation, structuring user-recognized behaviors
into narrative sequences, and (3) the use of meta-level controls to strategi-
cally undermine fragmentation of the agent's behaviors. Preliminary results are
encouraging, but further work, preferably involving the development of sup-
port for graphical presentation, will be necessary in order to fully evaluate the
implications of and possibilities for the architecture.
More generally, if black-box behaviorism involves thinking of human life
mechanically, reducing it to a matter of cause-effect, while narrative allows for
the full elucidation of meaningful intentional existence, then it seems likely that
narrative - and by extension the humanities, for whom narrative is a modus
operandi - can address meaningful human life in a way that an atomizing sci-
ence simply cannot. If humans comprehend intentional behavior by structur-
ingitintonarrative,thenAImustrespectandaddressthatwayofknowingin
order to create artifacts that stimulate interpretation as meaningful, living be-
ings. This suggests that the schizophrenia we see in autonomous agents is the
symptomatology of an overzealous commitment to mechanistic explanation in
AI, a commitment which is not necessarily unhelpful (since it forms the foun-
dation for building mechanical artifacts), but needs to be balanced by an equal
commitment to narrative as the wellspring of intentionality.
References
Baur, Susan (1991). The dinosaur man: Tales of madness and enchantment from the back ward .
New York: Edward Burlingame Books.
Blumberg, Bruce (1996). Old tricks, new dogs: Ethology and interactive creatures .PhDthesis,
MIT Media Lab, Cambridge, MA.
Brooks, Rodney (1990). Elephants don't play chess. In Pattie Maes (Ed.), Designing
autonomous agents . Cambridge, MA: MIT Press.
Brooks, Rodney (1997). From earwigs to humans. Robotics and Autonomous Systems, 20
(2-4), 291-304.