22 February 2011

The question of life can no longer be avoided because of Watson

In the 50s Alan Turing in Computing Machinery and Intelligence proposed what is known as the Turing Test, which measures whether a machine is truly intelligent. Basically a human asks questions of A and B and has to determine which one is a machine and which one is a human. No machine has successfully beaten the test.

Until IBM's Watson went on Jeopardy!.

Sure the voice could use some tweaks, but the Turing Test is purely a text-based test. That test Watson would win every single time.

Why is this significant? Because now we can no longer dance around the subject of when does a machine become a living thing. No more weird "qualia" talk. No more "does a soul exist?" Simply put, at what point of advancement do we consider machines to be living beings?

Yes, humans built and programmed Watson. While Watson obviously has the equivalent IQ of 150+ Watson is dependent upon humans. Watson is not alive. Not yet at least.

How do we define "living being?" Is it different from a virus? Before answering consider that viruses are capable of reproduction, have intelligence (whether measurable or not), have adaptive abilities and evolutionary processes, and quite possibly socialization.

Watson, in the correct environment could reproduce. Watson is capable of adapting. And possibly of evolution and socialization.

Is language a criterion? How do we then define language? Bees communicate but in no manner we truly understand. Watson communicates not only in English (and understands inferential references which are the most difficult to understand) but with various Machine languages.

Watson's one major limitation is its reliance on electricity. And that is the last defining line between non-living and living: self-sufficiency. Once Watson, or some other machine, is capable of self-sufficiency then machines will be Machines - an entirely new species.

No comments:

Post a Comment