Watch a TV series such as Battlestar Galactica, and ponder the future of intelligence. Is it indeed possible to create a “skinjob”, essentially an entity which is biologically similar in every way to humans except with the processing power of a multi-processor system, and enhanced senses? Can we encapsulate the workings of the human brain inside silicon? The truth is that it is very unlikely. It may be possible to meld the human mind with computers of some form, as in the Borg of Star Trek’s, but an autonomous intelligent entity? Maybe more like the hubots from the Swedish scifi show Real Humans.
Contemporary robots can do rudimentary thinking, they can walk up stairs, play chess, and soccer. But these are not revolutionary algorithms. Vision systems are used extensively in factory robots to search for defects in products, they are fast and efficient algorithms, but they are not truly intelligent. They follow a series of rules to process data. Sometimes algorithms “learn”, but they have to be given the data in the first place, they cannot search for it themselves, nor can they interpret the data the way the brain does. A case in point is the perception of beauty. Robots may one day have the intelligence to perceive and interpret a scene, the same way human vision works. However will they be able to look at a field of sunflowers and appreciate their beauty? Or peer into the heavens and wonder what lies beyond the stars?
Is it possible to make a machine that is more intelligent than a human? It may process things a couple of billion times faster, but is it smarter? Everything that an electronic brain does should be in the scope of a human brain because it is the human brain that designs the algorithm used by the machine. In the end, AI is not really AI. It is more likely some form of quasi–intelligent computability. For an object to have true artificial intelligence requires traits such as intuition and flexibility, to look beyond their programming. Humans are not given vast amounts of data to learn from… well that’s not exactly true. Humans are surrounded by data, which we process into information. We learn to associate the word “tree” with a woody plant, that can be deciduous or coniferous, but once we know what a tree is, our brains can identify the 100,000 species of trees by sight – and differentiate some of them based on size, shape, or leaf structure. A computer could do the same, but it has to be given information, it cannot infer it.
Besides which do we really need artificial intelligence? There are enough intelligent human and animals on the planet, that we don’t need to create it just for the sake of it. As HAL from 2001 put it…
“I know I’ve made some very poor decisions recently.”
Humans make enough poor decisions, we don’t need artificial intelligence to make them as well.