Over the last few months, the media have been positively bristling with so much hype about the coming AI Revolution as to make the heads of hundreds of millions spin. How it will completely upset the way things are produced and services, rendered. How it will increase productivity by anything between 40 and 1,200 percent (depending on whom you believe). How, it is “more powerful than Ukraine and Taiwan.” How it will upset the existing international order, assisting the US and India at the expense of China and Europe (Russia, apparently, does not count) and possibly saving the first-listed from what many people see as its imminent decline.
I am seventy-seven years old. I have never commanded a military formation, nor run a corporation, nor done research either in the natural sciences or in computing, AI included. In other words, my understanding of the issue is, let’s be charitable, limited. On the other hand, I feel that what little understanding I have of the way history works—an understanding I’ve been trying to acquire since I was ten years old—gives me some kind of handle when thinking about it. By way of a peg on which to hang my thought, I have chosen an article on the subject: S, Sharma, “8 Ways Artificial Intelligence (AI) Can Help You Improve Productivity.” The AI Journal, 9/1/2023, at https://aijourn.com/8-ways-artificial-intelligence-ai-can-help-you-improve-productivity/
So here comes Mr. Sharma’s list of some of the things AI is going to do.
1. “Forecast Demand Accurately.” As any student of economics can tell you, demand—here referring exclusively to commercial demand, not to every other kind—depends on many different factors. Technological developments, as when new gadgets, e.g steam engines or automobiles or computers, appear on the market. Prices, especially relative ones, that go either up or down. Macroeconomic developments. Changing circumstances, e.g droughts, global warming, or the emergence of epidemics such as COVID. Changing tastes, habits and ideas which cause the public to prefer one product over another. The discovery of new resources or the drying-up of old ones. The outbreak of war. Trends, luck, and fate (whatever that is). Some of these factors are foreseeable to some extent, others not. Some can be quantified and made computable, others not. All interact, forming a tapestry infinitely more complicated than anything that ever came off a loom. As result, for every correct vision there are ten incorrect ones. Briefly, the future is as much of a mystery today as it was 2,000 years ago when the Roman orator and lawyer Marcus Tullius Cicero discussed the problem with his brother Titus. Ironically Microsoft Bing, asked what the future would be like, first told me there were too many different scenarios to count and then invited me to submit my own.
2. “Automatic Text Creation.” This is already happening. Indeed multinational companies, in need of multilingual catalogues to sell their wares in different countries, have been using something like it for years. However, as anyone with experience in the matter can tell you, the outcome is likely to be both error-prone and moronic. Error-prone, because the machine has no idea of what the words it manipulates actually mean and is therefore liable to come up with all kinds of absurdities. Moronic because, not having an idea, it strings them together on the basis of the order in which they have been arranged before. One is reminded of the story in Gulliver’s Travels where “the most ignorant person, at a reasonable charge, and with a little bodily labor, might write books in philosophy, poetry, politics, laws, mathematics, and theology, without the least assistance from genius or study” by using a special table on which all the verbs, all the adjectives and all the verbs in all languages, written on rotating cubes, can be arranged into a text simply by turning a crank. And without in the least understanding what she (or he) has done, of course.
3. “Predict Maintenance.” With or without the aid of AI, predictions of maintenance requirements have been made for ages. At best, AI will make the process faster and more reliable. But I cannot help wondering whether AI will be able to come up with new and improved maintenance philosophies. If I am wrong, please let me know.
4. “Easy Data Extraction and Review.” This, too, has been done for ages. In fact every single ancient writing system known to us was designed specifically for that exact purpose. So in China, so in Mesopotamia, so in Egypt, so in the Aegean (Linear A), and so among the Inca (quipu). But whether General Motors e.g is better run today than it was at its heyday in the 1920s, long before computers and the current AI revolution, is doubtful. Why? Because demand adapts itself to the available supply. The more AI we have at our disposal the more complex and the more numerous the problems it is asked to resolve.
5. “Seamless interaction.” Attempts to achieve it, some successful, some not, have been going on for ages. The problem? Either changing external circumstances or the kind of human caprice known, euphemistically, as “the free will.” Or, not seldom, some more or less weird combination of both.
6. “Improve manufacturing processes.” This may be true, but it is certainly not new. As long as humans have produced anything—meaning, for the last few hundreds of thousands of years—they have also been trying to improve the production process. It worked: now slowly, now very fast.
7. “Automatic hiring.” AI may help employers go through the hiring process faster. However, often it does so only by making life for would-be employees much harder, as by having them fill far more questionnaires. All without any guarantee that things will work out better than they have done in the past or are doing now. For example, can anyone seriously argue that Julius Caesar was not as good in choosing his subordinates as George Marshal was?
8. “Social Commerce & Livestream Shopping.” This has been taking place at least since the first known cities, emerging some 5,000 years ago, set apart special spaces for—yes, you guessed it—“social commerce.” Indeed “commerce” itself stems from the Latin word cum, meaning “together.” AI may facilitated and extended it, but without changing anything really important.
Conclusion. In the words of Ecclesiastics “the thing that hath been, it is that which shall be; and that which is done is that which shall be done: and there is no new thing under the sun.” Mankind has survived the ice age when, if scientists got it right, at one point there were only 3,000-10,000 individuals left. It will survive AI too.