Artificial intelligence’s march of progress

Without doubt, artificial intelligence (AI) and machine learning (ML) systems have scored numerous remarkable successes over the past few years:

1. Defeating the reigning world chess champion.
2. Defeating two champion contestants on the American quiz show Jeopardy.
3. Defeating the world’s best human Go player.
4. Defeating all other entries in the 2020 Critical Assessment of Protein Structure Prediction (CASP) competition, a development described by German biologist Andrei Lupas as “This will change medicine. It will change research. It will change bioengineering. It will change everything.”
5. Demonstrating a weather prediction system that exhibits better accuracy than state-of-the-art conventional systems, at significantly less computational cost.
6. Generating impressive reports and analyses on diverse topics, notably with the recent emergence of tools such as ChatGPT and DeepSeek, which are remarkably more capable than even a few years ago.
7. Demonstrating notable success in finance, investment and planning.
8. Demonstrating astonishing facility to assist in mathematical and scientific research.

Technological progress through the ages

It is clear that we are witnessing a revolution that may well rank among the most significant in the history of science and technology, fully on a par with the printing press, the steam engine and the internet. As we will discuss in greater detail below, some observers of AI are predicting broad-based, worldwide economic advances dwarfing the current rates of progress, with perhaps as much as 20% or 30% annual growth in global output and a corresponding increase in global standard of living. This may well seem preposterous (and it does seem to the present author). However, as a July 2025 Economist article observes, for the vast majority of human history, so was the notion that the world economy could grow at all, or that there could be any significant change in overall standard of living from one generation to the next.

The Economist article observes that for the 17 centuries up to roughly 1700, global economic output grew at only 0.1% annually — essentially zero. Each generation could expect the same grim life of hardship and poverty as the one before. Global life expectancy varied between 25 and 30 years (as recently as 1880 it was only 35 years). Then with the advent of steam engines, spinning jennies and similar technologies, global growth in the 1700s quintupled to 0.5% annually by 1800. It then nearly quadrupled again by 1900, to 1.9% annually, then increased to 2.8% annually by 2000. Growth has not just continued; it has accelerated.

Along this line, 18th century economist Thomas Malthus predicted that population growth would soon outstrip agricultural output, resulting in dire poverty and mass starvation. This prediction did not fare well — world population did indeed rise, but various agricultural and processing advances led to even higher food production. Similar predictions were repeated in the twentieth century, notably with Paul Ehrlich’s The Population Bomb, but the ensuing green revolution, coupled with advances in processing, storage and transportation of agricultural products, so far has kept the human population fed quite well, even after duly acknowledging the deplorable pockets of poverty and war that persist. These developments, together with advances in medicine, computer technology (e.g., Moore’s Law) and numerous other aspects of modern life, have culminated in the current global life expectancy of 72 years, a figure that exceeds that of the wealthy nations of North America and Western Europe as recently as 1970.

In addition, these advances have had the remarkable and unexpected benefit of reducing fertility worldwide, thus reducing the pressure on future food production and per-capita economic growth. Analysts now predict (see HERE and HERE) that the world population will soon level off, perhaps as soon as 2050, and then decline. Numerous major nations already have fertility rates significantly less than the 2.0 required for population stability.

In short, although many significant challenges remain, environmental and social, the goal of feeding and providing at least a medium standard of living for the world’s teeming millions now seems achievable. The advent of AI may well accelerate these changes.

Nirvana or apocalypse?

Economists and others modeling the progress of technology have predicted that as AI-based technologies ideas generate more ideas more rapidly, these advances can compound, producing tools that are even more capable, generating even more rapid growth, leading to a technological “singularity” that we can only dimly imagine. In his book The Singularity Is Near (see also this 2011 Time article), futurist Ray Kurzweil predicted that this would occur by 2045.

Futurists such as Kurzweil certainly have their skeptics and detractors. Sun Microsystem founder Bill Joy has voiced concerns that humans could be relegated to minor players in the future, if not extinguished completely. But even setting aside such apocalyptic scenarios, there is considerable concern about the societal, legal, financial and ethical challenges of such technologies, as exhibited by the current backlash against technology, science and “elites” today.

Already there are numerous arenas of concern about the rapid deployment of AI:

1. Education: Numerous professors at universities large and small report rampant cheating by students using AI tools such as ChatGPT. Indeed, a recent study found that 90% of U.S. college students admit using ChatGPT or similar tools in their homework. Partly as a result, many professors have concluded that they can no longer base student grades on homework performance. Other teachers lament that many students, including some at elite universities, are unable to read and intelligently analyze any book-length assignment.
2. Loss of intellectual leadership: Along this line, many are concerned that the rise of AI will lead to “de-skilling,” namely the atrophy of intellectual acumen in an environment where ChatGPT and the like are doing all the real thinking: see for example HERE, HERE, HERE and HERE. This worry is not just a fancy. In a recent research study, several hundred U.K. participants were given a standard test of critical thinking and then interviewed about their use of AI chatbots. Younger users leaned more on AI but scored lower on the test. Another study found that after three months using an AI-based system to help flag polyps, physicians were less successful in finding them on their own.
3. Labor markets: Numerous large firms, including Amazon, JPMorgan and Walmart, have announced major white collar job cuts linked to AI. Anthropic CEO Dario Amodei recently warned that AI could potentially replace half of all entry-level white-collar jobs. Meta CEO Mark Zuckerberg says that most of the computer code for their Llama project will be written by AI within 18 months. Ford CEO Jim Farley has warned that AI will replace half of white-collar jobs. Other observers assert that this crisis is overblown, although they acknowledge that the transition will be difficult in any event.
4. Energy consumption and the environment: In the U.S., large data centers, driven in part by the demand for AI computation, already consume more than 4% of the nation’s total consumption, and this percentage is forecast to rise to nearly 10% by 2030. The rise of AI data centers is also raising significant environmental concerns, as their recent sharp rise in energy consumption is delaying the phasing out of older fossil-fuel generating plants by electrical utilities. Data centers are also being blamed for the recent rise in utility bills (perhaps unfairly), and noise pollution and other ills are concerns to those who live nearby.

When?

A key question here is when these potentially earthshaking events will occur. As noted above, some key figures in the field predict a major upheaval in the job market within just a year or two, as numerous large employers, both in the technology sector and elsewhere, announce layoffs or hiring cutbacks. So is the AI apocalypse already upon us?

Not so fast. To begin with, many of these same warnings are being cited in calls for new governmental regulations on the deployment of AI. For example, California has introduced a legislation that would place limits on the power that AI systems would be allowed to have over employees. The European Union has passed a ban on “high-risk” AI systems in schools and the workplace. The more visible these developments become, the more vocal will be the calls to slow down and tread more carefully.

Besides, there are limits to what currently envisioned AI systems can do. Suppose, for instance, that AI systems could replace 50% of the tasks in a certain organization. That leaves 50% that, for the foreseeable future, cannot be replaced. Besides, non-AI systems will continue to improve. Along this line, Hollywood studios have found it surprisingly difficult to incorporate AI technology into their production. Among other things, legal departments are fighting technical departments over copyright issues, such as how much human involvement warrants copyright protection. Partly as a result, no major studio has yet reported AI-generated films in development.

The history of British sailing ships provides a useful lesson. They were not rendered obsolete overnight with the introduction of steam-powered vessels in the 1850s and 1860s, for several reasons:

1. Sailing ships initially remained more economical for carrying bulk cargo because their propulsion (wind) was free, whereas steamships required substantial fuel (coal in most cases) and larger maintenance crews.
2. Early steam vessels were bulky and unreliable, so their initial use was limited to harbors and short shipping routes. The first transatlantic steamships carried sails as a backup.
3. Sailing ships continued to be used for long-distance, low-value freight such as wheat and corn.
4. The British Navy initially resisted steam power, fearing they would lose their sea dominance in the transition.
5. Twenty or more years elapsed between the first steamships and ships using advanced triple-expansion engines that were more powerful, reliable and economical.

Very likely the transition to AI will follow a similar trajectory, possibly compressed in time. It will be interesting in any event!