Nobel Prize winner: AI unemployment panic may become a "self-fulfilling prophecy" Silicon Valley should not overstate the threat of AI

Since the rise of the artificial intelligence (AI) boom in 2023, major technology companies have invested trillions of dollars, but public acceptance of this technology has been continuously decreasing. According to a recent survey by the Pew Research Center, only 16% of Americans believe that AI will have a positive impact on society in the next 20 years, while 40% of respondents expect its impact to be negative. People have many reasons for their aversion to AI, such as the disruptive issues caused by data centers and the large amount of water they use. But the most prominent concern is still that AI may take away jobs. Nobel laureate in economics Robert Shiller warns that this panic may ultimately turn into a "self-fulfilling prophecy". Shiller, who won the Nobel Prize for his research on how markets misprice risk, published a commentary this week titled "This 'apocalyptic thinking' must stop", further expanding on related theories. Shiller now focuses more on the reasons behind this mispricing and believes the key lies in narratives, the stories people tell each other about the future of the economy. He warns, When millions of people make millions of decisions based on pessimistic expectations, fear itself can create the reality people fear. Technology anxiety often precedes actual substitutes Concerns about new technology taking away jobs are nothing new. But Shiller points out that in every technological revolution in history, public fears have been far ahead of actual job replacements. In the 1830s, British textile workers protested against the power loom, a fear that was amplified by newspapers. In the 1920s, the play R.U.R. was a success, telling the story of robots revolting against their creators. Shiller also cites that the 1929 stock market crash itself was not enough to cause the Great Depression, as only about 2% of American households owned stocks at the time. What exacerbated the economic crisis was the sudden shrinkage in consumer spending, which stemmed from a general sense of uncertainty about future income. Similar incidents occurred after World War II as well - the 1957-1958 economic recession was labeled the automation recession by the media, blaming factory automation machines, but the economics community redefined it as a normal cyclical economic slowdown. Shiller worries that similar misjudgments are happening again. He writes that the slowing job market has many reasons, but reports have suggested that fears of mass unemployment caused by AI "are intensifying hiring freezes, and consumer confidence has fallen to historic lows." In March of this year, a survey by Quinnipiac University found that 70% of Americans believe AI will reduce human employment, up from 56% a year ago. However, research from Yale University's Budget Lab found that since the introduction of the chatbot GPT in late 2022, the most vulnerable occupational categories to AI have not shown significant changes. Calling on Silicon Valley to stop propagating the risks of AI unemployment Shiller believes that this round of panic is more severe because it is being fueled by leaders in the AI industry themselves. Dario Amodei, CEO of Anthropic, once said that AI could replace half of white-collar entry-level positions in the next five years; Mustafa Suleyman, head of AI at Microsoft, predicted that most white-collar jobs would be automated in 12 to 18 months. However, both later revised their timelines. The only ray of hope mentioned by Shiller lies in leadership. He cites a recent paper stating that U.S. President Franklin Roosevelt's fireside chats in 1935 significantly increased consumer spending among residents of cities with higher radio coverage. In conclusion, Shiller said, Washington's ability to change these narratives is ultimately limited. Not to mention that Donald Trump is not Franklin Roosevelt. "Therefore, perhaps we can only appeal directly to the tech leaders in Silicon Valley who are constantly propagating negative narratives. Emphasizing how powerful and even dangerous your AI models are may indeed help you sell more products, but if the economy enters a recession, this strategy may ultimately prove ineffective," he said. This article is reprinted from "Caishenhe". Editor: Chen Siyu.