AI's paradox: promise of abundance and fear of scarcity

Artificial intelligence is increasingly presented as the technology that may finally solve humanity's ancient economic problem: scarcity. Elon Musk has gone so far as to predict an "age of abundance" in which AI and robots make human labour largely unnecessary and money eventually loses much of its purpose. It is an extraordinary vision. But history should make us cautious. Technologies that solve economic problems often create political and social ones.

The Industrial Revolution offers an instructive precedent. Mechanisation enormously increased productivity. Agricultural productivity increased, transportation improved and industrial production outpaced previous eras. Yet industrialisation did not immediately produce prosperity for everyone. Alongside unprecedented wealth came overcrowded cities, dangerous factories, labour exploitation, unemployment among displaced artisans, colonisation and extraordinary concentrations of economic power.

The machine was not inherently unjust. The problem was who owned it, who controlled production and how the resulting wealth was distributed. It took decades of political and social struggle before societies developed institutions capable of managing the consequences: trade unions, labour legislation, universal education, progressive taxation, social security, workplace protections and eventually welfare states. AI may confront us with a similar challenge, only potentially much larger.

The optimistic argument is compelling. Combine increasingly capable AI with advanced robotics, and enormous quantities of work presently performed by humans could eventually be automated. Robots may construct buildings, manufacture products, cultivate crops, transport goods and provide services. AI could supply education, diagnosis, legal assistance, design, analysis and countless other forms of knowledge at negligible marginal cost. Humanity could become extraordinarily productive. But productive abundance is not the same as distributive abundance.

Suppose machines can produce ten times as much wealth as we do today. Who owns the machines? Who owns the algorithms, data centres, energy infrastructure and robots through which that abundance is generated? If productive assets remain concentrated among a relatively small number of corporations or individuals, society could become vastly richer while ordinary citizens become economically weaker. That is the paradox at the centre of the AI revolution.

More importantly, even where technology creates abundance, access to that abundance depends upon institutions. Food can exist while people remain hungry. Houses can exist while people remain homeless. The economic capacity to produce something does not automatically generate the political arrangements required to distribute it fairly. The central question of the AI age therefore may not be whether machines can produce enough, but whether ordinary people retain a meaningful claim over what those machines produce.

There is an additional problem. Work provides more than income. It provides identity, social interaction, routine, recognition and often purpose. If AI eventually separates employment from economic survival, humanity will have achieved something extraordinary. But we will simultaneously confront a deeply human question: what gives life meaning when earning a living no longer organises much of our existence?

None of this means the AI future must be dystopian. AI could improve healthcare, eliminate dangerous work, democratise knowledge and release human beings from much unnecessary drudgery. But history offers a warning. The Industrial Revolution demonstrated that technological progress and social progress are not the same thing.

AI may eventually solve many problems of production. It may even move humanity closer to material abundance. But unless societies answer questions of ownership, distribution, accountability and power, the technology that promises liberation could create new forms of dependency and inequality. The decisive question is therefore not simply how intelligent our machines become. It is whether our political and social institutions become wise enough to govern the power those machines create. At the moment, there is a mismatch between technological speed and human wisdom!