In his third Reith Lecture, Stuart Russell, Professor of Computer Science at the University of California, Berkeley, considers how developments in Artificial Intelligence technology will affect the wider economy. If, in the future, machines are doing the work that us humans used to do, what happens to us?
“It is now possible to envisage personal computers, small enough to be taken around in one’s car, or even one’s pocket. They could be plugged into a national computer grid, to provide individual enquirers with almost unlimited information.”Leon Bagrit
The machines are coming
Automation has been a thing since the ’60s, and Leon Bagrit, a British industrialist and pioneer of automation, was something of a visionary in this area. He predicted nano-technology and even the internet, saying ‘It is now possible to envisage personal computers, small enough to be taken around in one’s car, or even one’s pocket. They could be plugged into a national computer grid, to provide individual enquirers with almost unlimited information.’
“If every instrument could accomplish its own work… the shuttle would weave, and the plectrum touch the lyre, without a hand to guide them, chief workmen would not want servants, nor master slaves.”Aristotle
But fear of machines, or ‘technological unemployment’, as economist John Maynard Keynes called it, is not new. In the 19th century, radical textile workers, known as Luddites, smashed newly introduced mechanised looms, fearing that the machines would reduce their work. But, even as far back as the 4th century BC, political philosopher, Aristotle, predicted that “If every instrument could accomplish its own work… the shuttle would weave, and the plectrum touch the lyre, without a hand to guide them, chief workmen would not want servants, nor master slaves.”
Should we “be afraid of artificial intelligence!”?
Russell considers whether it’s right to fear technological advances. He posits that, in spite of the agricultural revolution, and the industrial revolution, work has always been available. He also proposed that innovation can bring opportunities; as AI continues to develop and ‘learn’, it will become possible for machines to master many more tasks that humans do.
There is also initially a ‘wealth effect’, when automation makes products cheaper. This means that we, as consumers, then have more money to spend on other products and services, thus increasing employment in another area.
Russell suggests that if AI is used intelligently it can save money that can be used elsewhere. Technology could be put to good use doing things that humans find boring, dangerous, or relatively expensive, such as removing graffiti, cleaning up the environment, inspecting shipping containers, or fighting forest fires.
Additionally, the loss of repetitive jobs, such as working on a factory production line, mightn’t be mourned much – just so long as those who lose their jobs could afford to pay the bills.
What happens to us?
But there are other problems. While technology can initially increase employment, by decreasing costs and increasing demand, there is a saturation point, meaning that fewer people will be required to do the work. There is also the evidence that low-skilled work has been paying less, in real terms, for decades. And what if those now unemployed people are not qualified or able to do a different job? What becomes of the taxi or HGV drivers once self-drive vehicles are rolled out?
And what if there is an unwillingness to plough back those savings into compensating decreased employment? New socio-economic arrangements need to be explored. One solution would be to provide a Universal Basic Income (UBI) to everyone, regardless of circumstance, allowing people to spend their time as they please. When Leon Bagrit anticipated the loss of work for humans, he also saw that it was important that people are educated to enjoy the fullest possible lives.
Where’s the money?
But who is really benefiting from new technologies? Government coffers certainly don’t seem to be filling at the same level. The rich are getting richer, and the poor are getting poorer. Amazon had sales income of €44bn in Europe in 2020 but paid no corporation tax.
The Guardian recently reported that “while British workers have suffered a near unprecedented squeeze in their wages, the richest thousand people saw their fortunes double in the first seven years after the financial crash. Covid has proved little different: Britain produced a record number of new billionaires in the pandemic.”
Thriving or surviving?
There are those who feel that UBI is an admission of failure, and that people wouldn’t ‘thrive’ if they didn’t have work. I’d counter that the ability to pay for one’s basic needs outweighs the loftier notion of whether or not a human is thriving. Surviving is uppermost in many people’s minds. Maslov’s ‘Triangle of Needs’ suggests that a human being’s most essential requirements are for shelter, food, and warmth. In the UK, these needs are not being met for a large number of people homeless and using foodbanks.
Russell considers that part of the problem is that we still don’t really understand what it is to be human. The care of our children, elderly and those less able to take care of themselves, are essential roles, and yet childcare and social care are currently badly paid and poorly regarded.
What’s the answer?
Whether we believe that the loss of work would be a positive or negative thing, Russell proposes that we need a radical redirection of science and education, both to equip individuals to ‘live wisely, agreeably and well’ and to support a human economy that would provide high-value interpersonal services. One of the most interesting positives about the future use of AI is within education. A child tutored individually can learn three times as much as in a classroom. If an AI system were developed to educate children individually it would be of enormous value to the world.
Of course humans would still be necessary to help with a child’s emotional needs, but this would fall within the range of ‘high-value interpersonal services’. It is these jobs, currently underpaid and undervalued like social care which cannot be done by AI that are likely to continue to be done by humans after other low-paid repetitive or dirty jobs have been taken over by robots.
As one of the audience pointed out in question-time, this is good news for feminists as currently most of these jobs are done by women !
How are we doing?
In 2018, the UK’s policy position on AI was fairly hands-off: “Existing sector-specific regulators are best placed to consider the impact on their sector of any subsequent regulation which may be needed.” More worryingly, the policy included an option of “Removing some existing regulatory burdens where there is evidence they are creating unnecessary barriers to innovation.”
The government also has a lot of catching up to do. An ‘Office for Artificial Intelligence’ was only proposed in 2018, and a ten-year ‘National AI Strategy’ was announced a mere three months ago. A ‘near £1 billion deal’ to boost the UK’s position in developing AI technologies has been promised, but it’s a drop in the ocean compared to the US and China, who lead the world in AI investment. Chinese AI companies raised a total of $31.7 billion in the first half of 2018, almost 75 percent of the global total of $43.5 billion.
And as far as the skills shortage goes, in 2021 the government proposed reforms to post-16 technical education and training, offering a flexible loan for higher-level education and training at university or college. Additionally a series of ‘Skills Bootcamps’ – free courses of up to 16 weeks – have been set up in collaboration with industry and universities. They cover a range of subjects, including AI.
Unfortunately a government focused on low taxation and less state intervention is unlikely to be able to afford the same investment as countries that are taking the development of AI more seriously, and, as Russell pointed out, our society will be the loser.