I recently met with the minister of a prominent church much concerned about the human condition in general and particularly on the impact of “artificial intelligence” (AI) software in displacing millions of gainfully employed people. Knowing that I was an investor in software companies, she asked me how I felt about being involved in an activity that was going to cause so much misery.
Of course, that question has been asked of technologists for some centuries since the Industrial Revolution produced machines that increasingly replaced human labor. Her question today is whether the new computer capabilities represent a dramatic increase in the ability of machine substitution of human labor. Are we entering a new age with a radical increase in the historical pace of industrial automation through robotics?
Many believe so. For example, the popular view is that increasingly intelligent computers will do just that, making human labor increasingly irrelevant. Much interest centers on self-driving vehicles where robots replace millions of commercial drivers.
A good place to begin an answer is to ask whether what we call computer intelligence resembles the human kind that it is supposed to replace. What is intelligence? ”Intelligence” is defined in the Merriam-Webster dictionary thus: “1a. The ability to learn or understand or deal with new or trying situations. b. The ability to apply knowledge to manipulate one’s environment or to think abstractly as measured by objective criteria. 2. mental acuteness.”
So how does a computer programmed with the most sophisticated AI software fare under these criteria? Consider the first criterion – dealing with new or trying situations. The computer “thinking” consists of accessing data that is matched to the situation that it is asked to act on. Finding the match is only possible within the limits of the program and the data from which it draws the actional steps. Anything “new” outside the program’s parameter and the database that is being accessed will baffle the machine. The idea of “abstract” thinking is not relevant to computer operation.
On the second criterion, “mental acuteness,” the relevant competitive characteristic may be the computer’s ability to respond more quickly to situations via programmed instructions than humans.
This short comparison between human and computer ‘intelligence’ centers on the ability to deal with events that are repetitive and predictable
This short comparison between human and computer “intelligence,” therefore, centers on the ability to deal with events that are repetitive and predictable. Hence, properly programmed computers can outperform human in such tasks. This is why robots have successfully replaced human operators in activities like manufacturing. And given that computers do not get fatigued with failing awareness, they outperform humans in repetitive tasks or in finding small parameters in a huge mass of data – like image analysis or seeking specific data from massive flies.
But computers fail in dealing with the unexpected. For example, consider driving a car in a controlled environment where the possible situations the vehicle faces are limited. Properly programmed, a computer will deal with the situations presented to the vehicle because the possibilities are limited. However, driving in dense traffic, a computer-controlled automobile is tasked with dealing with an almost unlimited number of situations, many ambiguous, where its data store has no equivalent. Then, the computer would have to choose from a choice of nearly similar situations and make a choice.
It is argued that human drivers are faced with such situations as well and given the number of road accidents, many drivers do a bad job of appropriate action. The claimed belief is that a computer will end up with better choices and hence reduce road accidents. Unfortunately, there is no way of proving that without extensive practical experience.
Hence the computer does not possess a feature in our definition of intelligence: “dealing with new and trying situations.” That feature cannot be programmed, and humans possess this quality in varying degrees.
After a lengthy discussion with my friend, where did I end up with an answer? “Indeed, we have better and better robots,” I said,” but they are still robots – machines that do predictable tasks better than humans. As the computing and data management of the machines continues to improve, more and more complex tasks can be performed by them. But they are not suited for dealing with the truly unexpected, nor do they have the ability to innovate answers to the unexpected situations and problems.”
Hence they will continue to displace human activity as technology improves. But we are still in the historical process of better machines doing more for humans, freeing humans from humdrum and mind-numbing repetitive activities.
In that manner robots have created enormous value and human satisfaction. That pace will increase with better software and computers. But the underpinning of the economic world is human creativity. And this robots will not substitute for.