A Bright AI Future

Photo by Jason Leung on Unsplash

I stumbled across an article on artificial intelligence (AI) by Andrew Mayne that really captures how I think about it. The author also explains the economics very clearly. So I thought I’d share. I highly recommend the article for those interested in the topic at all. But since you won’t all want to go read it, I’ll summarize key points here and add a few minor comments of my own.

Reason Article:

https://reason.com/2024/04/28/in-the-ai-economythere-will-be-zero-percent-unemployment/

Preliminaries

We’ve worried about AI for a long time. In my lifetime alone, you have 2001: A Space Odyssey¹, which was a 1968 film but very influential for a long time, where the computer HAL behaves like a human and turns against the astronauts. You have, of course, Terminator (1984)², where the artificial defense network, Skynet, turns against the human population. There are many more such concerns raised by science fiction writers before and after both these films. These loom large in our collective psyche, I’ll admit.

There is a more tangible concern that is really a protectionist/isolationist sort of concern. In a protectionist/isolationist sense, it’s along the lines of robots, computers, and AI will one day steal my job. That’s akin to worrying about foreign competition, be it from foreign imports (protectionism) or foreign workers (isolationism) in the economic sense.

The first is a security technology concern. It also raises the worry of “who controls the AI”, if anyone. Perhaps the scariest thought is that no one controls it. But, the two most likely controllers are governments or corporations. Of the two, I’d prefer corporations and a highly competitive market. But, I suspect “Web3.0” might offer some glimmer of hope and I’ll comment on other thoughts.

When ChatGPT³ went live, the world exploded with sudden concern, however. That’s certainly warranted in certain industries and for people in certain jobs. As a professor, I definitely struggle with how to handle students using ChatGPT, how to compete in a world with other professors who use it for coding and writing to produce newer, better academic papers to compete with me. It also competes with me as a teacher (the other part of my professor job) since students can learn using ChatGPT as a tutor or even teacher.

But, I believe, and this article argues forcefully, that the net effect is and will be overwhelmingly positive with more jobs, more opportunities and more human flourishing.

Mayne’s Reason Article Main Point: Efficiency and Expansion

The main thrust of Mayne’s article is completely right and is as follows: AI is a major, widespread innovation leading to serious efficiency gains across many jobs and sectors of the economy. Improved efficiency means, by definition, being able to do more with less. In business that means business can produce more with the same labor or produce the same with less labor. The same is true for people. If I can do more in less time, I can choose to either (a) spend the same time and produce more, or (b) produce the same and spend less time. As long as the choice is freely mine, I’m unambiguously better off in both cases.

Mayne then cleverly walks us through the process. It’s a pretty standard economics story of how development and expansion works, and he handles it brilliantly, I think.

He argues that improved efficiency for businesses means they both expand and earn higher profits. Some may focus on expansion, others on less expansion and enjoying higher profits, but overall both generally occur. As long as people and markets are relatively free, then other businesses will enter the market to capture some of the profitable business. Increased competition then leads to lower prices (and lower profits), making those goods and services more affordable for everyone.

Both these - the effect on existing businesses plus the new entrants - mean a massive expansion of business which means demand for a wide range of jobs and other inputs from other businesses. There’s a huge increase in demand for inputs which generates demand for jobs down the line through other industries as well.

Furthermore, as jobs and workers’ incomes expand and as the prices for these goods/services decline, there’s more income to be spent on other goods and services, expanding demand for other current businesses and calling new businesses, goods, services, and markets into existence. All that further increases demand for labor and inputs.

Finally, he argues that this is always the case with development. This is indeed the standard story any economist would tell in a classroom, for example, and pull thousands of historical examples from around the world as supporting evidence. But he adds that AI is particularly broad based, hitting a wide range of industries at once, and hence will likely have an even larger innovative, improved efficiency effect than most past innovations. He additionally suggests that AI complements robotization well, adding even further to the amplification of its effects. Casual observation suggests he’s right.

He says it here:

“AI and robotics will keep growing the economy, because they continuously increase productivity and efficiency. As the economy grows, there's always going to be a widening gap between demand and capacity. Demand for human labor will increase even when AI and robotics are superior and more efficient, precisely because there won't be enough AI and robots to meet the growing needs.”

Some of His Examples

He gives an early, conceptual example to support that commercial AI and robotics increase efficiency.

Link: “You use an industrial robot to weld a car because a human welder would take too long and wouldn't have near the precision. You use ChatGPT to help write a grant proposal because it saves you time and means you don't have to pay someone else to help write it.”

It’s a brilliant example because these are two areas where one can both imagine the technology being employed and imagine humans still deeply involved. He later points out that companies like General Motors (GM) which is a heavy user of robotics today employs approximately 163,000 people.

Even more powerful and supportive of his overall claims while also addressing the related concern that the only jobs remaining will eventually be high-tech, programming, etc., he then points out that…

Link: “Alphabet, Google's parent company, now has 182,502 employees; GM has 163,000. More people work for Apple (161,000) than McDonald's (150,000). Meta—Facebook's parent company—has more employees than ABC, CBS, NBC, and Fox combined (67,000 vs. 66,000).

And they aren't all programmers. At Microsoft, fewer than half the employees are software engineers. For a conglomerate like Amazon, the percentage is even less. Amazon has tens of thousands of people delivering packages, and Apple has human staff working in physical stores”

He also reminds that, this outcome would have been incomprehensible just 50 years ago. If you told someone that such companies would have more employees than GM, they wouldn’t have believed it. Of course, you would have had to try to explain what “social media” is and what Google does which would have been impossible to explain without: internet, personal computers, smart phones and all the other innovations.

This only supports his broader point: we cannot even imagine today what the economy will look like and be dominated by in 50 years, a time horizon he believes will be shortened with AI innovation. True enough, in my opinion. Try to explain to my 13 year old son what it was like when I graduated with my PhD in 2003 before apple had made the iPhone! He literally can’t imagine it. He literally can’t imagine not having uninterrupted Netflix and YouTube access 27/7 at home, in the car, in a restaurant, in the US, in Europe, …. That’s how much the world has changed in just the last 10 or so years. I remember seeing my first iPhone in 2008 at a party. I was shocked. Since Steve Jobs announced their creation in only 2007, Apple has sold 2-3 BILLION!

Comparative Advantage: I Must Quote This!

In the article Mayne brilliantly walks through the economic logic of this development. Some examples are conceptual, others are historic. He constantly reminds us that we cannot today imagine the world, economy, and many job categories that will come in the future due to AI and robotization.

And in doing all this, to explain the economic logic inherent in the whole story, he calls on the great David Ricardo and the Theory of Comparative Advantage. I don’t want to go all the way into it here. It underlies the basic theory of all trade, exchange, allocation of labor, how you allocate players on a sports team and so much more.

In short, it’s an explanation of why even when one person is better at everything than someone else, they still demand the labor or services of the other person. I’ll illustrate it with an example in an old Econ 101 textbook, hopefully people still know who Michael Jordan, the famous basketball player, is. The story is this: Michael Jordan is better at basketball and, presumably due to his great physical condition, mowing lawn than the neighbor’s 16 year old kid. Nevertheless, he would be willing to hire the 16 year old to mow the lawn, even though he can do it way better and faster himself because it allows him to spend more time on basketball. Again, he’s better at both but it still doesn’t make sense to do everything yourself. There are gains for both him and the neighbor kid if they engage in trade (mowing in exchange for money) even though one person is better at both.

The same will be true with AI and robots. We still need other people of lower skill. They will not all be replaced and the growth of the market will actually generate more jobs for low skilled workers, not less.

And here’s the quote I have to share here. Right after explaining Comparative Advantage and applying it here, he writes:

Link: “It's basic math, yet government economists will huddle around a conference room table arguing that you need to keep all production domestic while ordering out for a pizza instead of making it themselves—even if one of them happens to be a fantastic cook.”

Bwah ha ha. I just love that. And, my guess is that every economist reading this just giggled too. It’s a brilliant quote.

My Added Spin: Humans and General Equilibrium

If you followed this far, you should read his article. It’s great. It’s an easy read, and he has a ton of other great examples of jobs we couldn’t have imagined, and so on.

Of course, as an academic, I know there are some exceptions in history for specific innovations and specific industries, but I know of no exception for development overall. Certainly some jobs will disappear forever and new jobs will emerge we can’t imagine. The same happened going from horse-drawn carriages to automobiles.

The author doesn’t miss that of course, citing that “[t]wo centuries ago, 80 percent of the U.S. population worked on farms. If you told one of those farmers that in 2024 barely 1 percent of the population would work on farms, he'd have a difficult time imagining what the other 79 percent of the population would do with their time.”

There’s another reason - technically two reasons - that I personally don’t worry about an AI-robot world where no human has a job and we live in poverty, starving on the street. Fundamentally, the end aim of all economic activity is to serve human interests and, general equilibrium is real.

Humans are the alpha and omega of all economic activity. You cannot name a product that is not ultimately in the service of meeting human desires. One computer may tell another computer to tell the robot to shape some aluminum into a can that another computer tells some robots to put in boxes and ship on a truck all to eventually get me a soft drink to consume. We make military equipment to protect humans. We make roads for cars to transport humans to places humans want to go.

There is no end-product that isn’t in the service of satisfying human desires. And that, my friends, is where general equilibrium kicks in. If you un-employ humans, their demand for these goods will fall, decreasing demand for the robots themselves. There is no economy where humans are all gone and robots do anything.

There’s a popular concern about telling AI and robots to make as many paperclips as possible. First, that would be a stupid request because there is still a cost to each one being produced since robots use energy, wear their parts and need repair and so on. But let’s ignore that for a moment.

Suppose a business did this: maximize paperclip production. The abundance of paperclips would drive their price down near zero and that business would soon close. When it closes and can’t pay its electricity bill, the robots turn off.

Suppose AI and robots start leading to higher unemployment. That lowers incomes and decreases demand across the economy. Will your local restaurant really invest tons of money (and continuing maintenance costs) into the latest robot to supply pizzas if there are no customers due to mass unemployment? Or would they rather higher one of the now, very cheap unemployed people willing to work? As they de-robotize and hire people, jobs return and bring demand and livelihood back.

Students have literally asked me - after being scared by online videos and warnings - what happens if robots replace “all the workers” and humans have no more jobs. This is, to be impolite, the stupidity of the argument for a “universal basic income” as the solution. If no one has a job, the government has no one to tax to finance a universal basic income. (Slap forehead emoji :-) )

Anyway, none of those concerns keep me awake at any level.

Conclusion

Read the article. I hope you like it. I’ve thought about an article on AI and this guy did a better version than I could have, and speaks with more authority as a practitioner in this area.

While the long-run is bright, in my opinion, there can be challenges. To start, certain jobs will likely disappear. I might be in one of them as a university professor. So, we can think about that and people in such fields should definitely think about it and how to transition. Personally, I think I have some years and we might even need more professors as we all become more effective at doing our jobs. So, we’ll see.

Second, who controls the AI and robots is a concern. All concentrated power makes me nervous from big brother government to mega corporations. If I have to choose, I choose corporations. They don’t have guns and prisons and can’t make the laws that truly can coerce me.

Mega-corps and all that are still a problem. But the solution is strong rule of law, strong individual property rights and trying to keep as vibrant, free and competitive a marketplace as possible. The world is not and never will be perfect, but at least we can limit power and its abuse by making the corporations compete with each other to serve our human desires better.

And, recalling my argument that all these things are ultimately to serve human interests, demand grows for wresting control from both mega-corps (or any corps) and government both: decentralized networks, block-chain technology, and Web3.0 all promise to give more ownership back to the individuals the whole mess is designed to serve in the first place.

All this will have challenges we can’t imagine today, for sure. Could Arnold Schwarzenegger appear naked from the future and kill us all? I mean, I guess, anything can happen, but I’m not betting on it. Fundamentally, I’m optimistic and look forward to the exciting new world to come. If I happen to see Arnold tomorrow, I think I’ll just ask for his autograph.

Thank you for reading!

1. Wikipedia on the film: https://en.wikipedia.org/wiki/2001:_A_Space_Odyssey

2. Despite the belief of many, I WAS NOT the body double for Arnold in this field. I was only 14 years old, but I appreciate the flattering thought.

3. I use ChatGPT throughout to represent all generative AI of the ChatGPT type. There are now several.