The rapid advances in artificial intelligence in recent months have unleashed a tidal wave of worries. Will this new technology substantially reduce employment by eliminating the need for most human workers? Will it undermine democracy? Does it pose an existential threat?
Concern about technological change is nothing new. But it typically addresses what economists would describe as marginal effects: whether a larger share of workers without college degrees will find it slightly harder to get jobs, or whether income inequality will increase to some extent. Unease about AI, on the other hand, is of a different order of magnitude, with some experts predicting that it could upend civilization – or even wipe it out.
Why we should be more optimistic about AI
Tech leaders have argued that certain AI systems ‘pose profound risks to society and humanity’, a sentiment echoed by leading AI scientists. A recent YouGov poll found that nearly half of respondents are concerned ‘about the possibility that AI will cause the end of the human race on Earth’. Over two-thirds support a pause on some kinds of AI development.
This view is astonishingly pessimistic.
Let’s start with the basic – but seemingly overlooked – fact that technological advances improve human welfare. In 1800, 43 per cent of children died before the age of five. But in the following centuries, technological progress led to the development of drugs and therapies, new ways of treating disease, and productivity and wage growth. By 1900, around one-third of children died in their first years of life. In 2017, global childhood mortality was down to 4 per cent.
Moreover, agricultural technologies have boosted food production and preservation, reducing hunger, and advances in energy technology, such as electrification, have improved the lives of billions. Overall, technological innovation has reduced poverty by generating wealth.
Economists and other experts in the early nineteenth century could never have predicted the types of tasks that workers perform in today’s world.
To be sure, the process of ‘creative destruction’ unleashed by generative AI will eliminate the need for human workers to perform many of their current tasks. But the pessimists must remember that creative destruction creates as well as destroys.
Economists and other experts in the early nineteenth century could never have predicted the types of tasks that workers perform in today’s world. How could John Stuart Mill have foreseen that technological advances would one day lead to jobs such as systems analyst, circuit layout designer and fibre scientist? Imagine trying to explain Bruce Springsteen’s job to David Ricardo. There is no need to go back that far: the MIT economist David Autor and his co-authors found that the majority of current employment is in occupations introduced after 1940.
Similarly, concerns that AI poses a threat to democracy reflect undue pessimism. While ‘deepfakes’ – AI-generated images and videos that are synthetic but appear real – of political leaders and candidates could be used in sinister ways, new technologies also enable authentication of videos and images. Such tools are already being developed, and the financial rewards from meeting the demand for them will ensure that they remain reliable.
Changes for the better
In fact, democracy could be strengthened by advances in AI. One of the technology’s most promising opportunities is in education: AI applications could conceivably act as private tutors for every student. This should brighten the outlook for democracy’s long-term survival. As James Madison wrote, ‘Knowledge will forever govern ignorance: and a people who mean to be their own Governors, must arm themselves with the power which knowledge gives.’
Will AI wipe out humanity? It looks more likely to do the opposite. AI is already being used in drug development, including for Covid19 vaccines. A future pandemic – perhaps one much more lethal than Covid19 – might be stopped in its tracks by an AI-developed drug.
Like any powerful technology, generative AI should be appropriately regulated.
AI could also help scientists to gain a better understanding of volcanic activity, which has been responsible for mass extinctions, and to detect and eliminate the threat of an asteroid hitting the Earth. These optimistic scenarios seem more plausible than the pessimists’ view that AI could somehow use our nuclear weapons against us.
That is not to say that the ride won’t be bumpy. The rapid development of generative AI will disrupt labour markets, and the resulting economic turmoil will be painful for many workers. And it may take time before media and political leaders learn how to expose and shut down deepfakes.
Policymakers should not be complacent. Compared to past waves of automation, more should be done to help workers who face AI-related disruption. In some countries, it may be necessary to strengthen the social safety net. Like any powerful technology, generative AI should be appropriately regulated, with an eye toward ensuring its development is not unduly stifled.
The right solution is not to panic or indulge in undue pessimism. Instead, we should be reasonably confident that, like all general-purpose technologies before it – electricity, electronics, modern transportation, the internet – generative AI will improve human welfare. The specific changes AI will unleash in the economy and throughout society will be impossible to predict – but, on the whole, they will be changes for the better.