It is clear that with the opportunities created by artificial intelligence, the volume of content – articles, images, films, audio, etc. – on the internet generated by AI is exploding.
This gives rise to a paradox, namely that, for example, ChatGPT or Midjourney require content – text and images – to be trained on, and if more and more of the content in the world is created by AI, we end up with self-referencing AI, where we train future versions of ChatGPT on texts already written by previous versions of ChatGPT.
There is a parallel to this in economic theory, known as the Grossman-Stiglitz paradox.
In the article ‘On the Impossibility of Informationally Efficient Markets’ from the American Economic Review (1980), economists Sanford Grossman and Joseph Stiglitz challenge the Efficient Market Hypothesis (EMH), which essentially states that prices, for example, of stocks reflect all available information in the market, and therefore, one cannot ‘beat the market’.
It’s the ‘Wisdom of the crowds’. No one is wiser than the knowledge gathered in market prices.
But this creates a paradox, as highlighted by Grossman and Stiglitz because if the market is efficient – meaning that all available information is already reflected in the price, then there is no incentive to uncover information. In other words, you might as well give up trying to beat the market, but if everyone gives up trying to beat the market, then no new information is created.
And we actually see this in global financial markets. Increasingly, both professional and private investors realize that trying to beat the market is futile, and what we observe is that more and more investors become so-called passive investors – simply buying ‘the market’, for example, by purchasing an ETF on the entire US stock market rather than trying to ‘pick the winners’ and buy individual stocks.
This has become possible because it has become much cheaper to buy such ETFs relative to so-called actively managed investment funds.
And perhaps even more strikingly, more and more of the world is becoming efficient. For example, I have worked quite a bit on pricing soccer players, and I initially had an idea that there was money to be made if one simply used the right quantitative methods – à la Moneyball – to find soccer players in the global market who were ‘cheap’. But the truth is that it is much harder than one might think.
Today, it is much harder to find a 16-year-old boy from Northern Norway who turns out to be a global superstar that no one else has discovered. The reason is that the price of uncovering this information has collapsed.
Today, there is information available about soccer players and soccer matches all over the world. We can watch their matches on YouTube or see results and statistics for even the most peripheral leagues in the world. And it also means that the prices of soccer players will converge towards being efficiently priced.
The Role of the Entrepreneur
But despite this, the world has not come to a standstill. Even though markets – even for soccer players – have become more efficient, there are still many actors who try to beat the market every day.
These actors can be referred to as entrepreneurs.
In economic theory, the entrepreneur has two roles.
The first is the equilibrium-seeking entrepreneur, as described by the British-American economist Israel Kirzner. This type of entrepreneur identifies mispricing in the market – something that is too cheap or too expensive (‘Buy cheap, sell dear’).
This Kirznerian entrepreneur can easily be imagined as a robot that constantly scans the market for mispricing using algorithms – and it could easily be AI.
Such algorithms have been known in financial markets for at least 30 years, and a large part of trading in global financial markets today is done by such robots.
And such robots are now spreading to more and more areas on the internet. It is obvious that more and more news reporting will be created by AI. There will be things that can be done cheaper and faster with AI than what humans can do today.
But traditional economic theory, including Kirzner’s, is often static. It operates in a given world, but the real world is constantly changing. There is innovation. We come up with new things.
Algorithms have not replaced innovation. This leads us to another economist, Joseph Schumpeter, who spoke of “Creative Destruction.” And the creative element is the interesting part. The entrepreneur as ‘Gyro Gearloose’ – as the inventor or innovator.
The Gyro Gearloose-entrepreneur is not the one who buys and sells existing goods and services, but the one who invents new products and solutions.
The parallel to AI and content production is clear. The human contribution to content creation is precisely to be creative. To create something new. To get ideas.
AI cannot (yet?) replace this creativity. And for the world not to stagnate, human creative contributions are needed.
We need to embrace AI to fulfill the role of Kirzner’s entrepreneurs. This has already happened in financial markets, and it will happen increasingly in other areas. But at the same time, we need to embrace human creativity, the creative.
Many of the successful people I have met are often dreamers who are not limited by conventional notions. Although many of them realize that they are subject to the same conditions as everyone else, some of their ideas turn out to be groundbreaking.
These people are Schumpeter’s creative entrepreneurs. They are necessary in an AI age. Kirzner’s mechanical entrepreneurs are increasingly being replaced by algorithms and AI.
Finally, I must say that I have had the idea for this article for quite some time, and this morning I thought I would have ChatGPT write it.
Despite feeding ChatGPT with all sorts of inputs – and old articles that I and others have written – ChatGPT was unable to create a readable text on the subject.
Therefore, if we really want to ignite the AI revolution, we need to add human creativity.
And yes, ChatGPT helped me with the proofreading of this article (in fact it mostly translated it from an article I wrote in Danish and came up with the headline), and then you can see if the drawing created based on the text with Dall-E in ChatGPT is funny enough.
Lars Christensen lacsen@gmail.com, + 52 50 25 06. See more about how to book me for a presentation or lecture on the economics of AI here.
