Many tech entrepreneurs have long suggested that guaranteed income could cushion job losses due to artificial intelligence and automation. The latest and largest study on this idea was conducted by Sam Altman, the man behind ChatGPT.
Michael Dwyer/AP
hide legend
toggle caption
Michael Dwyer/AP
The rise of artificial intelligence has raised concerns that these technological advances could eliminate millions of jobs. Silicon Valley entrepreneurs have thought about this, too, and have long floated an idea to soften the blow: government financial aid, with no strings attached.
The first results of the latest and largest study on the impacts of free money are now available – research led by the man behind ChatGPT.
In 2016, Sam Altman, CEO of OpenAI, proposed funding an experiment in something called basic income. In a blog post that year, he said some form of national payment would likely be necessary as technology has eliminated more jobs while generating massive wealth for others. So he said it would be good to study what might happen if people received regular paychecks from the government.
“Are people playing video games or creating new things?” Altman writes. “Do people, without the fear of not being able to eat, accomplish much more and benefit society much more?”
Tech-related job losses aren’t his only motivation. Altman has cited progress toward ending poverty, writing, “I also believe that true equality of opportunity is impossible without some form of guaranteed income.”
There were a thousand different needs
It took a while for his free money experiment to come to fruition, and in the meantime, dozens of other experiments have been conducted. The idea has also been bolstered by the success of federal relief checks and other aid during the COVID-19 pandemic. But Altman’s study is much longer than most and covers a more nationally representative group in rural, urban and suburban areas.
For three years, 1,000 selected low-income people in Illinois and Texas received $1,000 a month. (A control group of 2,000 others received $50 a month.) Elizabeth Rhodes, research director at Altman’s nonprofit OpenResearch, began tracking their financial status as she enrolled them.
“One person just graduated from esthetics school, but they can’t afford a cosmetology license,” she says. “Another person just had their phone cut off. Another person just had a car accident and totaled their car, and they can’t afford another one.”
She said the needs were many and only money could satisfy them. Rhodes said this study, like many others, found that people spent the extra money mostly on necessities: food, transportation, rent.
“We’re seeing an increase in the number of people who are paying for their housing,” she said. “So a lot of people are living with other people and may be moving out on their own.”
Many also put money in the bank. The biggest increase in spending was actually to help family and friends.
The experiment faced an unexpected challenge: the COVID-19 pandemic hit early. That complicated the research, but also meant it took place during a sudden spike in unemployment. “The money gave more people the ability to make employment decisions during the most turbulent period in modern history,” says Karina Dotson, head of research and analysis at OpenResearch.
For example, the study found that the extra money allowed one woman to take a pay cut for a job with room for advancement, and now earn close to six figures. But this improvement in job quality was rare.
Overall, people who received cash payments worked slightly less (1.3 hours less per week on average) and their partners did the same. This includes some people who worked 50 or 60 hours per week at more than one job.
Participants also reported having more free time.
Dotson recalls a single father who worked at a restaurant. “When he found out about the money transfers, he told us he immediately went to his boss and wanted to cut back on his hours so he could spend as much time as possible with his 4-year-old son,” she says.
As for Altman’s question about people’s ability to create new things, the study did indeed reveal a greater interest in entrepreneurship. But it wasn’t until the third year of funding that some participants, mostly black, took steps to start a business.
At the same time, many people reported significant reductions in stress and food insecurity at first, but that faded after the first year. Researchers aren’t sure why. Rhodes also notes that in some cases, the extra money actually led to unexpected expenses. For example, some recipients were able to buy vehicles that later broke down and needed repairs.
The OpenResearch team plans to do more in-depth analysis of where people moved during the study (the most common reason participants gave for moving was to be in a better school district) and how the money affected children’s educational outcomes.
Altman declined to comment on the results so far. But the report says that in the debate over whether basic income contributes to citizens’ long-term prospects, “our results support both sides.”
Basic income proponents say it is not a magic bullet on its own.
Guaranteed income is a long-standing idea that has a surprisingly diverse fan base, ranging from libertarian economist Milton Friedman and President Richard Nixon to the Rev. Martin Luther King Jr. and the Black Panthers. Other Silicon Valley billionaires who have backed it include Elon Musk and Jack Dorsey.
The most ambitious vision has been a universal basic income, as when 2020 presidential candidate Andrew Yang called for giving every American adult $1,000 a month (plus cost-of-living increases), regardless of income. In a 2016 blog post, Altman called for giving people “enough money to live on.”
But thinking about basic income has changed dramatically. Recent experiments and national policy proposals are much more limited and targeted at low-income households.
“I hope people take away from this study and other studies that a guaranteed income alone won’t work,” said Facebook co-founder Chris Hughes.
He also co-founded the Economic Security Project, which advocates for a basic income. But Hughes says it’s not a magic bullet: $500 or $1,000 a month isn’t enough to overcome the skyrocketing costs of housing, health care, education and child care. Still, he says, the growing body of research, as well as pandemic-related payments, shows that a little extra can help families survive.
“I think a good starting point would be to have a guaranteed income in case of hardship,” he says. To that end, Hughes suggests automatic payments could be set up when a rise in unemployment signals a recession.
But turning no-strings-attached cash into national policy would face strong opposition. Some states have even banned it.
“Contributing to society through the labor market … is a more promising system than one in which the poor simply receive a check from the government,” says economist Michael Strain of the American Enterprise Institute.
Although research to date shows a limited impact on employment, Strain is concerned that a permanent basic income program could exacerbate the long-term decline in employment rates for some groups.
A better idea, he says, would be to dramatically increase tax credits for low-income workers. For example, imagine a person loses a $40,000-a-year job because of automation, and the only other job they could find pays $25,000. “What if we lived in a world where the government gave you the $15,000?” Strain asks. “You only get it if you take the job. But the government will try to give you enough money to make it worth it.”
For the record, neither Strain nor Hughes are particularly concerned about massive job losses caused by technologies like AI. They argue that history shows that over time, new technologies create new types of jobs. But they agree that as employment becomes increasingly precarious, struggling American families need more help, one way or another.
