{"text":[[{"start":7.84,"text":"It is rare for a central banking institution to model the economic impact of human extinction (spoiler alert: GDP goes to zero). But a startling chart depicting that scenario was shown in a recent research paper from the Federal Reserve Bank of Dallas."}],[{"start":25.79,"text":"Forecasting the likely impact of artificial intelligence on US economic growth, the researchers presented three scenarios. Their central forecast was that AI might boost the trend growth of US GDP per capita to 2.1 per cent for 10 years. “Not trivial but not earth shattering either,” the report’s authors, Mark Wynne and Lillian Derr, wrote."}],[{"start":50.97,"text":"But the bank also considered what might happen if AI achieved the technological singularity, when machine intelligence surpasses the human kind and becomes ever smarter. "}],[{"start":65.1,"text":"In a good case, that superintelligence could trigger a massive rise in GDP and end scarcity. In a bad one, it could lead to the rise of malevolent machines and end humanity. There was, the authors noted, little empirical evidence behind either of these extreme scenarios, although some economists have been exploring both possibilities."}],[{"start":90.22999999999999,"text":"Evidently, there is a wide spectrum of views among economists about AI. But the economic consensus is that it might be no more consequential than some other technological advances, such as electricity, the internal combustion engine and computers. "}],[{"start":null,"text":"
"}],[{"start":109.11999999999999,"text":"It takes a massive technological jolt to shift an economy the size of the US above its growth trend line of just under 2 per cent a year. For more than a century, that trend has held pretty steady in spite of two world wars, the Depression and periodic global financial crises, not to mention myriad previous technological advances. "}],[{"start":133.35999999999999,"text":"Moreover, economists argue, the rapid diffusion of AI we see today will not automatically translate into productive adoption. Indeed, it may temporarily lead to a loss of productivity as jobs are shuffled and new ways of working are introduced — the so-called J-curve."}],[{"start":153.70999999999998,"text":"But AI evangelists hear such arguments with slack jaws. Many of them depict economists as a downbeat and conservative tribe, vainly trying to predict the future by looking in the rear-view mirror. The way they see it, automating brawn triggered the Industrial Revolution and automating the brain will lead to an even bigger jump in productivity. That should surely shift the trend line in a dramatic way."}],[{"start":183.85999999999999,"text":"Last week, the Stanford Digital Economy Lab hosted a seminar to debate the contrasting views of economists and technologists. The discussion was led by Tamay Besiroglu, co-founder of Mechanize, an AI start-up that wants to enable “the full automation of the economy”."}],[{"start":205.41,"text":"One way of thinking about AI, he said, was that it would enable us to inject significant new inputs into the economy by massively increasing the number of digital workers to tackle many more tasks. “AI effectively turns labour into a type of capital,” Besiroglu said. "}],[{"start":224.45,"text":"Another way in which AI might boost productivity is by generating and diffusing more useful ideas. In his great book The Enlightened Economy, the economic historian and recent Nobel laureate Joel Mokyr argued that the Industrial Revolution occurred in Britain when it did because of the rapid circulation and adoption of practical knowledge. The Dallas Fed paper acknowledged that AI could well accelerate discovery and innovation in unpredictable ways that would meaningfully contribute to higher living standards."}],[{"start":262.14,"text":"Although the differences between economists and technologists appear stark, Erik Brynjolfsson, director of the Stanford Digital Economy Lab, says they are not incompatible. “I think they both have a lot of truth to their positions. And there’s a way to reconcile them,” he told me."}],[{"start":282.78,"text":"After studying productivity gains from previous general-purpose technologies such as steam engines, electricity and IT, Brynjolfsson suggests the biggest economic impact often comes from investments in complementary areas, rather than from direct investments in these technologies themselves."}],[{"start":305.97999999999996,"text":"So, for example, it took a generation before redesigned factories were built to exploit the full benefits of electricity. He suggests the gains will arrive a lot quicker from “amazing” AI but they will still not be immediate. “These complementary investments are where the real action is. And they take time and are very complicated,” he says."}],[{"start":330.51,"text":"That suggests both economists and technologists may be partly right but precisely wrong. Productivity growth may eventually be far bigger than most economists currently forecast but far slower than many technologists predict."}],[{"start":357.22999999999996,"text":""}]],"url":"https://audio.ftcn.net.cn/album/a_1762828962_5819.mp3"}