The Covid-19 pandemic may have killed 2.7m people and resulted in the worst economic contraction in a generation. But the optimism of the West Coast tech elite remains undimmed. “The future can be almost unimaginably great,” as Sam Altman, who heads an artificial intelligence research company, wrote in a recent essay, Moore’s Law for Everything.
Like many tech evangelists, Altman argues that we are on the brink of an AI-induced productivity explosion that will shower abundance on all. The entry of millions of Chinese workers into the global labour force over the past three decades will be seen as nothing compared with the arrival of tireless AI “workers” that will radically cut the cost of all tradable goods. When robots start inventing better robots, we will achieve “Moore’s Law for everything”, not just computing power.
“Imagine a world where, for decades, everything — housing, education, food, clothing, etc — became half as expensive every two years,” he enthused.
Altman admits that many readers may think his arguments utopian. Others will dismiss them as those of a techno-crank. But it is hard to deny that Altman has an interesting vantage point from which to peer into how the future may unfold.
For many years, Altman helped run Y Combinator, the start-up accelerator that backed many of Silicon Valley’s smartest businesses, and he is now chief executive of Open AI, the research company that has developed the spookily convincing GPT-3 language generation model.
More unusually among the Silicon Valley crowd, Altman has long expressed a deep concern about the flip side of technological creation: societal disruption. As he sees it, the AI revolution will further shift the balance of power from labour to capital, resulting in the concentration of phenomenal wealth among the machine owners. Good societal outcomes will only result if there is a radical change in public policy and a massive redistribution of wealth. He has himself sponsored an experiment with universal basic income in Oakland and proposes the creation of an American Equity Fund to give everyone a share of future technological riches.
“A great future isn’t complicated: we need technology to create it and policy to fairly distribute it,” he wrote.
There are three main criticisms of Altman’s equation of AI + UBI = utopia. The first is that he wildly overestimates the short-term impact of AI. The second is that he wildly underestimates the difficulties of policy change. The third is that by focusing on the fascinatingly improbable, we will crowd out more useful discussion about the worryingly probable.
Many of those who work in the field of AI themselves think that AI is overhyped. Erik Larson, author of a forthcoming book called The Myth of Artificial Intelligence, is one. He is critical of many of the “false narratives” around AI which suggest that machine superintelligence is somehow imminent. AI systems that can soar in one narrow domain still have a hard time jumping off the ground in others.
For example, IBM’s Watson may have won the quiz show Jeopardy! But it failed to cure cancer. A few years ago, autonomous cars seemed to be just round the corner. But they still have difficulty distinguishing between a turning bus and an overpass. “There is an odd cognitive dissonance between the reality of what we are doing and the science fiction debates you see about AI,” says Larson. What worries him more is the “spectre of a jobless future from mindless automation”.
From the policy perspective, it is equally simplistic to believe that UBI is the remedy for all ills, although it may be a partial cure. Vi Hart, who also once worked at Y Combinator, acknowledges the danger of what she calls the “AI-pocalypse” but does not think that UBI is the answer.
“I believe the apparent synergy of AI and UBI comes not from them being a radical brilliant solution, but from them being the easy way out, the status quo, the default choice that leads further down our current path of increasing inequality,” she wrote.
Altman frames the AI debate as one of techno-determinist inevitability: the technological revolution is unstoppable and so it will be the fault of politicians if we fail to adapt. But this is buck-passing of epic proportions that gives technologists a free pass to develop whatever they want.
Their efforts would be better spent figuring out how best to use AI to enhance human creativity and innovation in small, discrete and meaningful ways, rather than render it redundant.