The research participants were split into two groups: ‘cash recipients’ and ‘the control group’. The recipients received a thousand dollars per month within three years. The control group remained uninvolved. They then tracked the difference of people’s social behaviour and adaptation.
OpenResearch, a social research non-profit, has published the results of a major unconditional cash study, which Sam Altman commissioned it to deliver earlier. The study itself was part of ChatGPT promotion programme: ‘how to preserve social stability when AI deprives ordinary people of their incomes’.
Research design: the participants were split into two groups: ‘cash recipients’ and ‘the control group’. The recipients received a thousand dollars per month within three years. The control group remained uninvolved. They then tracked the difference of people’s social behaviour and adaptation. Both groups balanced just a bit over the American poverty line. In a nutshell, they ‘filled up’ about a third on top of recipients’ regular annual income. So they could not just quit their jobs out of the blue, even if they deeply wished to do so.
Certainly, publishing the outcomes, the researchers were bragging about the drastic changes, which happened to the recipients of unconditional cash in comparison with the control group, and how ‘good and favourable’ all this was. They supposedly began taking care of their health while working less to spend more time with their families and plan their careers better. In evidence, to support the figures, they published interviews with respondents happily telling about their new life.
But if we focus on figures, not the press release itself, everything is completely different. The unconditional cash recipients’ increase in ‘private’ time was 1.3 hours per week. And if we look at social groups, this pertained to single mothers only. Of course, it’s cool when single parents get a chance to spend more time with their children thanks to a higher income, but this is obvious without experiments.
At the same time, if we deduct ‘helicopter money’, the recipients’ own earnings decreased. The motivation to chase the dollar has gone, as you are still in the black.
The group’s average spending to ‘care more of their health’ increased by a huge 20 dollars. Being aware of prices in the U.S., it is obvious a few of the group’s members had to spend the whole sum they got from the experimenters on their health. When split among all of them, the result was 20 bucks.
The only thing, which the extra income seriously impacted, is the level of stress as an integrated indicator for many things at the same time: from alcohol consumption to family fighting. This indicator has reduced by about a fourth, which is 10 times less than the average 2–5 per cent for all other ‘positive changes’. The fact that a higher income gives more confidence is another ‘Captain’s Obvious statement’.
The recipients started to eat better (in money terms), spend more on cars, some even managed to improve their living conditions and save a bit more. But no one has reached major, life-changing results, as far as I understand.
One cannot say the experiment was useless though. It confirmed that Marxists were right speaking about the class structure of societies and the economic basis of classes’ existence.
While you are inside your class, you live like everyone else and whether you get an X or an X+20 per cent does not really make a difference. Well, yes the top part of the class in terms of their income eats better and cuts down less on gasoline. And that is pretty it.
The substrate for social change forms when there is a dramatic adjustment of incomes. If the experiment participants had received three times they were used to, we would have seen a funny bunch of scenarios. It would have been the same when the experimenters had taken from them two thirds of what they used to get.
Overall, Altman’s conclusion is clear: if the AI revolution keeps the incomes of the poor and the middle class, even if it cuts them a bit, no one will notice it. And there will be no social impacts at all. But how to achieve this if the base scenario and the prerequisite of a successful AI revolution is ‘the creation of vast wealth owned by a few with technological unemployment for others’?