Special Session
Selmer BRINGSJORD, Rensselaer Polytechnic Institute (RPI) and Micah CLARK, NASA
What Recent Developments Bring Us Closer to HAL 9000?
In an age wherein reasonably intelligent
and apparently credible
people say that we are on the verge of The Singularity, the event that
brings into existence machine intelligence that very rapidly exceeds
that possessed by our species, we submit that it's interesting to
consider the talk's title with some care. While we expect discussion
and debate at P&C to range over many aspects of this question,
in our
prepared remarks we specifically consider a specific variant of the
general question, viz.,
*What recent developments bring us closer to HAL 9000's apparent
ability to detect, counter, and produce lying?*
Under this variant, we (i) consider formal definitions of lying
suitable for instantiating in a computing machine, (ii) present
evidence that a "lying machine" can be engineered, and (iii) discuss
what additional evidence and engineering would be required to get all
the way to not only HAL 9000, but even smarter less vulnerable HALs on
NASA's upcoming mission to Mars. Finally, (iv), by showing relevant
portions of the *2001* movie itself, we explain that while people have
thought about HAL's lying, they haven't though much about his
*lie-detection* -- despite the fact that this is "in" the movie.
David G. STORK, Chief Scientist of Ricoh Innovations
2001: HAL's Legacy
Stanley Kubrick and Arthur C. Clarke's
1968 epic film "2001 - A Space Odissey" included of the most compelling
and thoroughly researched visions for computer science ever depicted in
film, specifically the HAL 9000 computer.
This presentation will compare the visions
in the film with actual developments in computer science, all in the
namesake year. What did the film´s creators "get right" or
"get wrong"? Why?
You will never see the film the same way again.