Something has been rubbing me the wrong way about the term “artificial intelligence”.  I don’t think  the adjective “artificial” fits well with the noun “intelligence”.  Some background thoughts…. Whatever we define intelligence to be, which is a subject of much debate, I believe it is a property that something either has or does not. There is of course our perception of intelligence, which I consider to be a separate problem.  If something fits our definition of intelligence, then our ability to perceive that intelligence has little to do with whether or not that thing actually possess it.

So if something is intelligent, how can we say that it is artificial…or “not real”.  It could be non-human intelligence, or non-biological intelligence but that doesn’t make it artificial.  Intelligence is a property that something has regardless of how it emerges.  If something truly displayed intelligence, how can we call it artificial?  We don’t even know how intelligence emerges in biological organisms, so how can we characterize those same properties emerging in some other system “artificial”?

To me that would be like saying a fish is truly wet… but if I dump a bucket of water on a person, then the person is artificially wet.  The thesis being that the fish is wet “naturally” and the person is not.  Does it matter? Both are still wet.  The fish is no more wet than the person, regardless of how they got that way.  Being wet is a state of existence.  The mechanism by which you became wet can be characterized, but the end state itself is no different.

To that end, I think the term artificial intelligence is just not a good fit.  It is first and foremost presumptuous, identifying anything other than biological intelligence as inferior.  As mentioned before, we don’t know how biological intelligence works, so who is to say some other kind would be better or worse?  Second, if we are talking about technologies that we have built to solve problems, to imitate biological behaviors or to explore biological intelligence, then I think the term simulated intelligence, or partial intelligence would be more appropriate.  The distinction be that we are recognizing that what we have produced is a mere shadow of what we consider intelligence to be, rather than position it as true intelligence that is artificial.  The former both acknowledges that we have not arrived at constructing true non-biological intelligence while simultaneously not intimating that if we did ti would be second class.

I doubt I will at any time in the near future be on a crusade to change our terminology, I am sure I will use the term AI, but secretly (I suppose not so secretly anymore), I will cringe a little.