Post date: 16-Feb-2012 23:45:40
I use the term "Intelligence Engineering" to describe what I would like to do, instead of the widely used term Artificial Intelligence. I am interested in building genuine intelligence rather than the artificial kind.
For the most part, this is just a playfully disrespectful and arrogant distinction. It is disrespectful of the scores of Artificial Intelligence researchers who are far smarter and more productive than I. It is surely arrogant to suggest that I will do better. Results in both Artificial Intelligence and the greater Cognitive Sciences are both daunting and inadequate.
However, the notion of artificial intelligence is an odd one. It seems to me that a device must be either intelligent or not, or at least intelligent to some degree. The famous Turing Test has perhaps unwittingly underpinned the term artificial intelligence as it prompted people to distinguish between 'genuine' human intelligence and that other 'artificial' kind masquerading as human. From this perspective, the term artificial intelligence may appear to be founded in a homocentric prejudice and the Turing test might be seen as a prompt for a definition of Intelligence.
I also use the term Intelligence Engineering to make it clear that I have aspirations in Engineering rather than Science. However the two activities are clearly related and are decidedly blurred with regard to Intelligence. I doubt that "Intelligence" will yield to formulas as clean as those in physics or algorithms as clean as those in computer science. Here in particular, "the proof is in the pudding", and in a sense, the science is the engineering.
"The real proof is in the pudding."
Rumelhart and McClelland