MIT Economist: Markets are overpriced for AI-driven productivity

Economists, consultants and business leaders are eager for anything that will boost productivity growth in industrialized countries. It has been disappointing in the information age, despite all the glory and talk of revolutionary technology. Total Factor Productivity (TFP) – economists’ favorite measure of economic productivity that estimates how much aggregate output is growing as a result of improvements in labor and technology – used to grow by 2 % per year throughout the 1950s, 60s and early 70s. Since the 1980s, its growth has been hovering around 0.5%. The promise of AI-driven productivity is music to everyone’s ears.

It’s not just wishful thinking on the part of businesses. The hype machine of the tech world is powerful. We are told every day in newspapers and social media about the results of the revolution of new devices, sparkling with superhuman intelligence.

And of course, the prospect of artificial general intelligence (AGI) fascinates us after decades of Hollywood movies where machines become so capable that they fight humans.

Alas, it seems impossible that anything of the scale promised by the world of technology – such as the rapid development of a singularity where machines can do everything that humans do they can do it – it’s possible even remotely. Even more extreme estimates such as Goldman Sachs’s that artificial AI will increase global GDP by 7% over the next decade and from the McKinsey Global Institute that the annual growth rate of GDP is may rise by 3-4 percent between now and 2040, to be overly optimistic.

What should we expect from AI?

My research shows that the impact of AI technology may be in the range of about 0.5%-0.6% increase in US TFP and about 1% increase in US GDP over time 10 years old. This is no joke. Given the state of the economy in the United States and other industrialized countries, we should welcome such a contribution with open arms and do everything in our power to realize this potential. However, it does not change.

Where this number comes from is good to understand, not only to increase our confidence in it but also to know why we can destroy that ability if we give in to protest.

In its current state and with current capabilities, the biggest impact of AI will come from automating some jobs and making workers more productive in other jobs. Currently, this can only happen in jobs that don’t involve a lot of interaction with the real world (construction, maintenance services, and all kinds of blue-collar and manual jobs) and in jobs that don’t have a role. the core of society (intellectual, mass entertainment and academics are out). Even in jobs that fall outside of these categories, it will be difficult to get productivity growth from AI. Doctors can benefit from AI to evaluate and measure their treatment and physician decisions. But this requires more and more reliable AI models – not such visionaries as the giant speech models that can write Shakespearean songs.

Based on the available evidence and these facts, I estimate that only 4.6% of jobs in the US economy will be affected by AI in the next decade.

Combine this with existing estimates of how much business can benefit from the use of AI-powered tools, which is around 14%, and you come up with a TFP boost of just 0.66% over ten years, or 0.06 % annually.

I immediately agree that there is a great deal of uncertainty. It is possible that AI product types will advance significantly in the next few years and may suddenly make more than the 4.6% I currently estimate. Or they could revolutionize the scientific process, leading to new methods and products that we could not even dream of today and completely changing the manufacturing process for the better.

But I, for one, don’t think this is a likely outcome. A very small percentage of US companies are currently using AI, and it will be a slow process until AI is used economically.

Hype is the enemy

Worse, hype can be the biggest enemy of increasing productivity from AI, and the misallocation of resources it causes can cause us to lose the modest gains we can get from AI.

This is due to at least three reasons. First, with the hype, there will be more investment in AI. Most business executives, at least until last week’s market correction and gasping for air, were under pressure to jump on the AI ​​bandwagon. If you don’t invest heavily in AI, you’re lagging behind your peers, journalists, consultants and technologists have been told. This leads to a loss of performance and not to the benefit of efficiency. In the rush to automate everything, even processes that should not be automated, businesses will waste time and energy and will not receive any of the promised productivity benefits. The hard truth is that achieving productivity gains from any technology requires organizational change, significant additional investment, and upskilling of employees, through training and on-the-job learning. Amazing, revolutionary miracles from AI may remain a chimera.

Second, there will be a lot of wasted resources, investment and energy, as technology companies and their sponsors pursue larger and larger AI product models. The current market correction will not stop technology leaders from asking for billions of dollars to buy more GPU capacity and strive to build larger models. They may pass on some of these costs by selling their services and technologies to businesses that are not ready to make this transition, but as a society, we are truly bearing the fruits of this overinvestment.

Third and most important, increasing productivity requires workers to be more productive, acquire greater knowledge, and use better information when making decisions and solving problems. This doesn’t just apply to journalists, academics and office workers—most of what electricians, plumbers, blue-collar workers, educators and health workers do is deal with a series of of problems. When they use better information, they will become better at their jobs and will be able to perform more complex tasks. The real promise of AI is an information tool: collecting, processing, and presenting reliable, contextual, and easy-to-use information to decision makers.

But this is not the way the tech industry, enthralled by chatbots and human-like dreams of AGI and misled by self-proclaimed AI prophets.

Other comments worth reading published by Good luck:

Opinions expressed in Fortune.com’s comment sections are solely the opinions of their authors and do not necessarily reflect their opinions and beliefs. Good luck.

Suggested information: CEO Daily provides key news that leaders need to know from across the business world. Every weekday morning, more than 125,000 readers trust CEO Daily for information about—and from inside—the C-suite. Register Now.