Technology is an absolutely huge part of our world. Virtually everything we accomplish, from mundane tasks like household cleaning to exploring the farthest reaches of outer space, is done with the indispensable aid of machines of our own creation. Cleaning chemicals, computers, space probes, submersibles: these are all bits of manmade ingenuity that can propel us to new heights and new depths. Yet all of these technologies are fundamentally unintelligent.
I call the vast majority of our technology unintelligent because it requires constant human supervision to act in a nonrandom order, to perform the work it was designed for. Cars today are staggeringly complex and remarkably powerful yet still are reduced to a mere piece of scrap metal without a human operator. The burden of negotiating even a simple straight road rests not on the car but on its driver.
Examples of this absurd reliance on human operation are found everywhere in every field. Doors require humans to open them and lights require humans to turn them on. Computers are powerful yet require humans to perform much of the higher-level work. Without a human giving very nuanced commands a computer would not have a clue what data to process and how. It seems today that everything has buttons or controls of some sort, and without constant monitoring of those controls the whole of our technology is useless.
Indeed, our technology is powerful because it is specialized. Every piece of equipment is designed to perform a specific task. However, this idea of creating disjointed modules working together not only wastes time due to the huge amount of human control needed but also results in unreliability. In the large majority of cases where technology has gone wrong, the human operator is to blame. And our technology is not remotely resourceful or alert enough to even notice there is an error.
Even the BP oil spill can be attributed to nothing other than human error. This article cites a memo that lists a variety of warning signs that were ignored or misinterpreted in the time leading up to the spill. In operations of this scale much can go wrong, and small errors are to be expected. Yet conventional technology does not expect errors and cannot adapt to them. And when humans fail to make the necessary responses the technology can fail altogether causing a debacle like this one.
Significant research is being done to make our technology more adaptable, more flexible, and more intelligent. This is one focus of the field of artificial intelligence. A huge deviation to explore the myriad directions of research in AI is not necessary here. I only wish to show a few examples of technology that isn’t stupid.
Problems to which AI is typically applied are of the sort that I have identified above. Specifically, it addresses tasks that are often relegated to humans. These include pattern recognition, language processing, image processing, and diagnosis. This last item is particularly relevant to our discussion; it means the process of determining whether an error is present, where in the system an error is located, and what kind of error it is. This strikes at the core of my main criticism of current technology.
Yet this sort of system is far from being everyday. It is used more commonly in some fields (especially where data processing is relevant) than others, but there is no question that practical AI is still a young and rapidly developing enterprise. For the time being, though, what we have in our technology is AS – artificial stupidity. This is my criticism: our cars, planes, and computers rely on constant human operation for both routine activities and adaptations to change.