Has the artificial intelligence bubble burst? This is the question on many people's lips this week as we witnessed the stock market crash, wiping out billions of dollars in the value of technology companies that have boomed since the advent of large-scale language models. It was economist Hyman Minsky who believed that Wall Street was over-incentivizing risk and tricking investors into parting with their cash for the wrong reasons. The result was a devastating boom-bust cycle. His model started with “displacement,” which means investors get excited about things they don't fully understand, like AI.
Last year, OpenAI's ChatGPT chatbot reached 100 million users in two months. According to Minsky, what follows after displacement is boom, euphoria, profit taking, and panic. There was a boom. It's reported that more than 200 AI startups around the world are unicorns, or companies worth more than $1 billion.
There has been plenty of enthusiastic reaction. Earlier this year, The Wall Street Journal reported that OpenAI CEO Sam Altman was in talks with investors to raise $5 trillion to $7 trillion to restructure the global chip industry and ensure the supply of necessary computing power. The premise here is that all we need to get closer to a sense of life is more computing power and more access to the data needed to process it. Surely the expectations are too high for this bet to be a sure success.
It's entirely possible that we're in a profit-taking phase. This is when the stock market goes through a Wile E. Coyote situation: the cartoon character falls off a cliff, looks down and realizes he's standing on nothing, and plummets. In the market, investors suddenly realize that prices can't keep rising, feel the need to pull out their cash, and stock prices crash. This is happening with semiconductor stocks, specifically Nvidia, the only company currently capable of making the graphic processing unit chips that power AI.
Whether we've reached the Minsky panic stage is up for debate. Bubbles pop just as quickly as investors flee the market. Amazon's shares plummeted in part because analysts looked at the company's earnings and realized it had invested lavishly in AI without much to show for it. Some venture capitalists think the party hasn't even started yet. But so far, the public hasn't found much use for the technology. A Reuters Institute poll found that 29% of Britons said they had used ChatGPT, but only 2% use it daily. Analyst Benedict Evans writes that most people who tried chatbots “didn't see how it could help them.”
While these concerns may plague the market, humanity has a much bigger problem. Many argue that if the development of AI is not carefully managed, humanity risks extinction itself. Politicians say the technology is advancing so quickly that the public needs to be protected. California, in particular, is considering a bill that would hold not only those responsible but also the original developers of AI models liable for misuse, putting many lawmakers at odds with the state's major industry. Social considerations will surely be more effective at stopping the rapid growth of AI than financial speculation.
