format_quote Originally Posted by
Serinity
What is the ruling on AI?
One problem will certainly be that "artificial intelligence" is a contradictory term.
If there exists an algorithm for a computer device to carry out a calculation, this calculation does not require intelligence. Generally spoken, if a computation can be executed by a computing device, it means that the computation does not require more than an underlying (Turing-complete) instruction set. A contemporary CPU with support for arithmetic, logic, and branching will by definition suffice.
These basic instructions are utmost simple. They are not considered "intelligent". A young child can learn to do them. A computing device just does them very fast. A predictable composition -- no matter how complex -- of such simple basic instructions, would in my opinion still not be intelligent. The other way around is true too. If the composition of such instructions is not predictable, it will also be impossible to test. It will therefore also not be a valid program for execution by a computing device.
For example, "computer vision" is considered by some people to be "artificial intelligent". The problem is that the implementation of computer vision software (such as
opencv) mostly rests on
wavelet mathematics proposed by
Alfred Haar in 1909, and that there are good reasons to object the practice of calling applications of mathematics "artificial intelligence". That is really a misnomer. Mathematics are not some kind of "artificial intelligence".
On the other hand, software implementations that have no basis in mathematics are always snake oil. In that case, we are not dealing with "artificial intelligence" but with clever attempts to mislead you into believing that it is.
There are many things that computers cannot do, and some (but not all) of these things are considered "intelligent". If we manage to make these computers do these things, it means that we came to understand their mathematical basis. This certainly happens once in a while. However, mathematics will never be "artificial intelligence".
In fact, "artificial intelligence" is mostly a nebulous term that cash-strapped academic researchers use in order to obtain speculative funding from gullible government funding agencies.