The BS Detector: Artificial Intelligence?

This is a followup to my blog on the "Theory of Evolution of Intelligence" where I mentioned the most primitive useful form of artificial intelligence is the bullshit detector machine. Years ago we looked forward to the day when a computer could beat a chess master. We thought, if that ever happened we would be on the threshold of artificial intelligence. Some doubted that it could ever be done. Well, that benchmark has been blown away, so what's next? Let's skip all the games now and get to something of real use.

One possibility might be a machine that could drive a car. Although it would be a useful test of intelligence, unfortunately, this machine would only add to our woes as a species, since populations would continue to increase as the resources depleted.

It seems to me we really need is a machine that could actually reverse the backward slide of our civilization's sum knowledge. We already have a very primitive prototype of a BS Detector: the spam filters. The best of these filters include a look up on the Internet to see if this piece of spam is from a known spam source. Also the better ones can read a picture to recognize spam-like content printed on a photograph.

My proposed benchmark of a basic BS detector should be to input an article from a source such as the Globe and Mail, or a program segment from CNN or Fox News, and get an answer back indicating level of bullshit on a scale of 1 to 10. Where we could increase that scale to 11 in the case of some BS-rich sources.

The BS detector would need to separate the content as opinion or fact. If fact, the machine would check the sources against the Internet. If opinion, to evaluate the logic, reasonableness and probability.

This machine is just a theoretical machine at this point, and I don't know if Homo Sapiens has enough time left in its evolutionary run to actually build a working one, or if it's even theoretically possible. But think of the uses it might have. First, to avoiding a stock market crashes or economic depressions by feeding in market analysts blurbs, or even company reports. It would have been very useful to run a couple of AIG annual reports through a BS detector and have it come up all 11's on the BS scale.

The one problem is with the question of the universal bullshit scale, one that does not vary from one country to the next, or from one political party or religion to another. We would need a universal knowledge base, for example do we choose Wikipedia or Conservapedia, or one of the other variants that spring up because Wikipedia is seen as biased by some. At any rate, a machine, much like a human, needs some underlying base for reference. Producing a logical machine is simple enough, but knowing which sources to trust for facts is a problem that may be impossible to solve. My own bias would be to go with the scientific community consensus, but not everyone would trust that. So in the end we may have the Scientific-based BS Detector, and another Jesus-Centered BS Detector, and yet another North Korean BS-Detector and so on. Which of course opens us up to the possibility of networking the BS detectors together in a kind of debate or "fight to the finish".

A working universal BS detector would not actually be making advances in science for us, but it would increase our sum knowledge. That's because
(sum knowledge) = (all that we know) - (all the BS).
Eliminate the BS, our sum knowledge increases. And we are better able to move forward and solve some of our other problems such as famine, resource depletion, war, and overpopulation. Or to invent an even better artificial intelligence machine.