Charlie Craven, The Independent
Not for the first time the world is faced with revolutionary technology that threatens to change the make-up of the workforce. During the 1950s, 60s and 70s, millions of people found their jobs replaced by newly automated labour. The question loomed large: which jobs were safe? How long till my job is replaced? Who benefits from this automation? In 1971 philosopher John Rawls invited his readers to partake in a thought experiment. In A Theory of Justice he asked people to design a society, the rules, the laws, its economy and social structure, with the caveat that once you enter that society, you would have no idea what role you would play.
How would you design taxes, or benefits, if you had no idea if you were to be rich or poor? Would you structure policing or the courts differently if you had no idea about your race or gender?
It’s through this “veil of ignorance”, argued Rawls, that we can arrive at the most just society. Underpinning this concept is something Rawls called “distributive justice”. This was his model for deciding how to fairly share the wealth produced by a society. So, say you have a productivity of one, and I also have a productivity of one. By working together we can produce something more than the value of two. Through cooperation, there is a productivity that is greater than the sum of its parts, and can be harvested for the benefit of all. When a company of a few hundred staff work together they can produce billions, but it’s dependent on the custodial staff as much as it is on the CEO. A society working together is exponentially more effective than an equal number of individuals, but that productivity only exists because we are all working together.
The question that Rawls asked his readers was, what do we do with this extra productivity? Do we build hospitals, schools and roads? Or, is it vacuumed up into the personal bank accounts of a tiny minority?
Rawls was writing in a turbulent time. Since the start of the 20th century, the American economy had exploded with mass industrialisation, culminating in several cycles of boom and bust and skyrocketing living standards for almost everyone in the country. But by the late 1960s, this was starting to change. Rawls observed a society where the automation of industrial processes resulted in huge unemployment, while productivity continued to increase. The extra cash did not end up in the hands of the people who made it possible: the teachers, the street sweepers, the people who grew the food that fuelled the people who designed the new systems. It ended up in the hands of a growing number of billionaires, while middle and lower-income Americans saw their living standards fall. To remedy this Rawls advocated radical – for the time – positions: high-quality, state-funded education for all, a robust minimum wage, free healthcare. With the oncoming AI revolution, we face a similar period of disruption today. Just as when machines automate physical labour, algorithms are set to automate intellectual labour, and some sectors will lose jobs. Artists, writers, coders, lawyers and teachers all risk replacement under powerful new AI systems. But while the human element may be removed, the productivity coming from these sectors is only set to grow. Like Rawls, we must ask ourselves what is going to happen to all this extra productivity. There have been countless articles decrying the dangers of AI. Doomsayers warn about everything from new forms of misinformation to a Skynet-style extinction of the human race. But few writers (and fewer politicians still) have grappled with the more mundane, and more likely result of AI. That the spoils will be hoovered up by a few, while the society it disrupts will be left wanting.
If we are going to mitigate the biggest disruption that AI has to offer we need to engage in some radical thought experiments of our own. How would you design the society of 20 years’ time, after AI has become commonplace, if you had no idea of your race, gender, age or education? Or if you didn’t know if you were going to be a tech billionaire or have your job taken by a machine? Is universal basic income inevitable? Could we capture some of the AI-generated productivity to fund free social housing for anyone who needs it? If we see an uptick in wealth inequality, do we need to radically reorganise and refund higher education? If Rawls were with us, he might argue that AI itself is only possible because of the collective pull of everyone in society, and if we want a fairer society we need to build institutions that benefit everyone.