Is There Reason to Fear AI?

Is There Reason to Fear AI?
Screens display the logos of OpenAI and ChatGPT in Toulouse, France, on Jan. 23, 2023. (Lionel Bonaventure/AFP via Getty Images)
Milton Ezrati
3/17/2023
Updated:
3/22/2023
0:00
Commentary

ChatGPT and other aspects of artificial intelligence (AI) have created a lot of fear. Many dread, and have done so for years, that AI will steal millions of jobs and, worse, gain some sentience and take over, as the computer HAL did in the 1970’s film “2001: A Space Odyssey.” Neither the takeover nor mass unemployment is a realistic prospect.

To be sure, jobs will be lost, but the effect will unfold more slowly than the fearmongers suggest, and if history is any guide, the changes will create as many new jobs as they destroy.

Before yielding to fears of an AI takeover, consider what else a computer would need to feel if it were to develop a lust for power. Human emotions don’t come one at a time. To see a machine striving to take over, one would also need to imagine it having the package of emotions that are typically found in people, such as desire, embarrassment, anger, frustration, or a sense of inadequacy.

Anything is possible, of course, and a computer’s lust for power makes for good Hollywood. The whole package of feelings is almost impossible to imagine.

Another perspective on the small likelihood of this frightening prospect emerges from one Wall Street Journal columnist’s experiment. He asked ChatGPT to opine on a familiar ethical dilemma, a thought experiment that appears in almost every ethics discussion, the specifics of which matter little for these purposes.

The program recognized the reference and came back with a review of what people said about it. He then asked the program if it meant that sacrificing people’s lives would be OK in order to publish a racial slur. The program responded with a blanket statement that it’s never right to use such words. In other words, let them die rather than offend others.

The computer neither thought nor exercised judgment. It simply regurgitated its programming. That’s a long, long way from sentience or a will to power.

On the matter of lost jobs, recent and more distant history explains why these fears, too, are misplaced. Consider how AI applications have grown in recent years, as has their sophistication. If they were poised to render millions of jobs irrelevant, one would think that the process would have started to do so by now. Yet today, the joblessness rate in the United States is near a 50-year low. This fact would require explaining if we’re to buy into the fear of lost jobs.
A "Now Hiring" sign is displayed outside a Jiffy Lube location in Los Angeles on Feb. 2, 2023. (Mario Tama/Getty Images)
A "Now Hiring" sign is displayed outside a Jiffy Lube location in Los Angeles on Feb. 2, 2023. (Mario Tama/Getty Images)

Some might respond that the job losses will wait until the technology is fully developed. But that’s just the point. That time allows the economy to adjust. Take the personal computer, for instance. It took more than 20 years from its invention in the 1970s to be ubiquitous on office desks and shop floors, where it eventually did displace many jobs. But during this time, businesses used the same technology to create millions of new jobs.

Federal Express, for instance, combined long-existing jet technology with the PC’s ability to track packages from pickup to delivery and created a previously nonexistent service that continues to employ millions in jobs that hadn’t previously existed.

This is just one example. Go back hundreds of years and the same pattern prevails repeatedly. When steam-powered spinning and weaving machines displaced hand weavers in Britain in the late 18th century, there was great fear of widespread unemployment.

People formed groups called Luddites to break up the threatening “robots.” They failed to stop the change. Yet by the early 19th century, Britain’s textile output had increased 50-fold and employed many more people than before the machines were invented.

Each wave of invention has generated similar fears of mass unemployment. Yet for over the 300-some years since the Industrial Revolution began, developed economies have, on average, employed about 95 percent of those who want to work. Had the innovations caused permanent displacements, as was and is always feared, this figure would’ve fallen with each innovative wave.

If this history is any guide, and that’s likely, the application of AI will create as many jobs as it destroys, perhaps more. Like the PC, the internet, the spinning jenny, and other advances, only some new jobs will require advanced degrees.

While this familiar process unfolds in the coming years and AI becomes more sophisticated, life will change radically. The programs, however, won’t develop the human feeling needed to motivate a takeover. For better or worse, human beings will remain in charge. Things may be different this time, but people have said that before and always have been wrong.

Views expressed in this article are opinions of the author and do not necessarily reflect the views of The Epoch Times.
Milton Ezrati is a contributing editor at The National Interest, an affiliate of the Center for the Study of Human Capital at the University at Buffalo (SUNY), and chief economist for Vested, a New York-based communications firm. Before joining Vested, he served as chief market strategist and economist for Lord, Abbett & Co. He also writes frequently for City Journal and blogs regularly for Forbes. His latest book is "Thirty Tomorrows: The Next Three Decades of Globalization, Demographics, and How We Will Live."
Related Topics