How to solve AI’s inequality problem
In an essay called “The Turing Trap: The Promise & Peril of Human-Like Artificial Intelligence,” Erik Brynjolfsson, director of the Stanford Digital Economy Lab, writes of the way AI researchers and businesses have focused on building machines to replicate human intelligence. The title, of course, is a reference to Alan Turing and his famous 1950 test for whether a machine is intelligent: Can it imitate a person so well that you can’t tell it isn’t one? This goal has been pursued by many researchers since then, according to Brynjolfsson. He says that the obsession with imitating human intelligence has led AI and automation to often replace workers rather than extend human capabilities and allow people to do new tasks.
Economist Brynjolfsson says that simple automation can lead to greater inequality in income and wealth. He writes that the excessive focus on AI-like human beings drives down wages while increasing the market power of a few who control and own the technologies. He argues that automation is the “single greatest explanation” for the rise in billionaires, at a time when many Americans are seeing their average real wages fall.
Brynjolfsson is no Luddite. His 2014 book, coauthored with Andrew McAfee, is called The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies. He believes that the thinking of AI researchers is too limited. He says that he has spoken to many researchers and they all agree that the goal of AI research is to create a machine that is “human-like.” He adds that it’s also a low bar.
In the long-term, however, he believes AI can create far more value than just replacing workers. He says that it’s easier for businesses to simply swap in a machine than to think about investing in technology that uses AI to expand their products and increase the productivity of their workers.
Recent advancements in AI have been remarkable, resulting in everything from driverless cars and human-like language models to AI. However, it is crucial to guide the technology’s trajectory. The choices made by researchers and businesses so far have resulted in new digital technologies creating enormous wealth for those who own them and inventing them. However, they have also created many opportunities for those in vulnerable jobs that could be replaced. While these inventions have created good tech jobs in a few cities like San Francisco or Seattle, much of the rest has been left behind. It doesn’t have be this way.
Daron Acemoglu, an MIT economist, provides compelling evidence for the role automation, robots, and algorithms that replace tasks done by human workers have played in slowing wage growth and worsening inequality in the US. In fact, he says, 50 to 70% of the growth in US wage inequality between 1980 and 2016 was caused by automation.
This was largely before the explosion in AI technologies. Acemoglu is concerned that AI-based automation could make things worse. Early in the 20th century and during previous periods, shifts in technology typically produced more good new jobs than they destroyed, but that no longer seems to be the case. Pascual Restrepo and his colleague believe that companies are choosing to deploy “so-so technology” which can replace workers, but little to increase productivity or create new business opportunities.
At the same time, researchers and businesses are not seeing the potential of AI technologies to improve the capabilities of workers and deliver better services. Acemoglu identifies digital technologies that could help nurses diagnose illnesses more accurately and teachers give more personalized lessons to students.
Government, AI scientists, and Big Tech are all guilty of making decisions that favor excessive automation, says Acemoglu. Federal tax policies favor machines. While human labor is heavily taxed there is no payroll tax for robots and automation. He also said that AI researchers are “no compunction” [about] in working on technologies that automate jobs at the expense of many people losing their jobs.
But, he reserved his fiercest ire at Big Tech, citing data that shows that roughly two-thirds AI work is funded by US and Chinese tech companies. He says, “I don’t think it’s an accident that there is so much emphasis on automation when technology’s future in this country is in control of a few companies such as Google, Amazon, Facebook and Microsoft that have algorithmic automation in their business model.”
Anger over AI’s role in exacerbating inequality could endanger the technology’s future. In her new book Cogs and Monsters: What Economics Is, and What It Should Be, Diane Coyle, an economist at Cambridge University, argues that the digital economy requires new ways of thinking about progress. She writes that “whatever we mean when we say the economy is growing, or things getting better,” the gains must be shared more evenly than in the past. “An economy that is dominated by tech billionaires, gig workers, and tech millionaires, will not be politically sustainable.” Coyle says that digital technologies will be more important to improve living standards and increase prosperity in many sectors, including construction and health care. People can’t be expected or expected to accept the changes if they don’t see the benefits, or if they only see good jobs being lost.
Coyle stated that she worries that tech’s inequalities could make it difficult to deploy AI. She says, “We’re discussing disruption.” “These are disruptive technologies that change how we spend our time every single day, and that change business models that work.” She says that social buy-in is necessary to make such “tremendous” changes. Coyle says that resentment is simmering among many because the benefits are perceived as going to elites in a few wealthy cities.
In the US, for instance, during much of the 20th century the various regions of the country were–in the language of economists–“converging,” and financial disparities decreased. Then, in the 1980s, came the onslaught of digital technologies, and the trend reversed itself. Many retail and manufacturing jobs were eliminated by automation. A few cities had new, well-paying tech jobs.
According to the Brookings Institution, a short list of eight American cities that included San Francisco, San Jose, Boston, and Seattle had roughly 38% of all tech jobs by 2019. New AI technologies are particularly concentrated: Brookings’s Mark Muro and Sifan Liu estimate that just 15 cities account for two-thirds of the AI assets and capabilities in the United States (San Francisco and San Jose alone account for about one-quarter).
The dominance of a few cities when it comes to the invention and commercialization AI means that wealth disparities will continue to rise. This will not only lead to social unrest but also, Coyle suggests that it could prevent the development of AI technologies necessary for regional economies to grow. Part of the solution may lie in letting go of the monopoly that Big Tech holds on the AI agenda. This will likely require increased federal funding for research that is not controlled by tech giants. Muro and others suggested substantial federal funding to support the creation of US regional innovations centers ,. A more immediate response is to expand our digital imaginations to envision AI technologies that not only replace jobs, but also expand opportunities in the areas that different parts of the country are most concerned about, such as education and health care.
The fondnesss that AI and robotics researchers have for replicating the capabilities of humans often means trying to get a machine to do a task that’s easy for people but daunting for the technology. For example, making a bed or a cup of coffee. Driving a car. It’s amazing to see an autonomous car navigate the streets of a city or a robot serve as a barista. Too often, those who create and deploy these technologies don’t consider the potential impact on employment and labor markets.
Anton Korinek is an economist at the University of Virginia. He is also a Rubenstein Fellow at Brookings and says that the tens of millions of dollars spent on autonomous cars will have a negative impact on labor markets once they are deployed. This will result in the loss of many drivers’ jobs. He asks what if those billions were invested in AI tools that could increase labor opportunities.
When applying for funding at places like the US National Science Foundation and the National Institutes of Health, Korinek explains, “no one asks, ‘How will it affect labor markets?'”
To support MIT Technology Review’s journalism, please consider becoming a subscriber.
Katya Klinova, a policy expert at the Partnership on AI in San Francisco, is working on ways to get AI scientists to rethink the ways they measure success. She says that AI research is tied to comparing or matching human performance. AI scientists evaluate their programs in image recognition based on how easily a person can identify an object.
Klinova states that such benchmarks have influenced the direction of the research. She adds that it’s not surprising that automation and even more powerful automation have emerged. “Benchmarks are super important to AI developers–especially for young scientists, who are entering en masse into AI and asking, ‘What should I work on?'”
But benchmarks for the performance of human-machine collaborations are lacking, says Klinova, though she has begun working to help create some. Collaborating with Korinek, she and her team at Partnership for AI are also writing a user guide for AI developers who have no background in economics to help them understand how workers might be affected by the research they are doing.
“It’s about changing a narrative away from one in which AI innovators are given a free pass to disrupt and then it is up to society and the government to deal with it,” Klinova says. Klinova says that every AI company has an answer to questions about AI bias and ethics. However, they are still not able to address labor impacts
The pandemic has accelerated digital transformation. To replace workers, businesses have resorted to automation. The pandemic has shown us the potential of digital technology to expand our capabilities. They have given us the tools to create new vaccines, and made it possible for many people to work remotely.
As AI grows in its power, it will be interesting to see if this causes more harm to good jobs and creates more inequality. Brynjolfsson says, “I am optimistic that we can steer technology in the right direction.” He adds that this will require us to make deliberate decisions about the technologies we invest in and create.
“The Turing Trap: The Promise & Peril of Human-Like Artificial Intelligence”
Daedalus, Spring 2022
“The wrong kind of AI? Artificial intelligence and the future of labour demand”
Daron Acemoglu and Pascual Restrepo
Cambridge Journal Of Regions, Economy and Society, March 2020
Cogs and Monsters: What Economics Is, and What It Should Be
Princeton University Press
I’m a journalist who specializes in investigative reporting and writing. I have written for the New York Times and other publications.