Blog: The surprisingly simple relationship between AI, global inequality, and your bank balance
Every product on Amazon might get cheaper. But at what cost?
Having reached new heights in humanity’s ancient pursuits of chess and Go, AI is now set to disrupt a beloved modern past-time: shopping. One of the most meaningful impacts of automation is likely to be a significant drop to the cost of goods – making many (if not all) things cheaper to buy.
The economic logic predicting that prices will drop is relatively simple: automating activities currently done by humans improves performance, usually due to higher speed, higher precision and the elimination of human error. In other words, taking humans out of the factory and replacing them with robots results in significant productivity increases, making it cheaper and faster to make goods.
The productivity gain and greater efficiency coming from automation means not just lower prices but increased economic growth, at least for the developed economies leading the technological charge. A related consequence is that wages should rise in those economies. So far, so rosy!
Unfortunately, there is a dark side to this apparent economic miracle. What might be an economic boon for wealthy countries may herald economic catastrophe for developing states. The potential advantages of automation are historic, but so too are the risks: it is no exaggeration to say that long-term global inequality is at stake.
The Robots Really Are Coming
Why should developing economies be fearful of automation? Continuing advances in artificial intelligence and robotics threaten to totally disrupt the economic model underpinning their development. Developing states disproportionately rely on the exact industries that (because they involve repetitive/manual work) are most vulnerable to automation. For instance, McKinsey has suggested that more than two thirds of manufacturing work done by humans could be done by machines – saving well over a trillion dollars in labour costs.
Simply put, if it is no longer the cheapest option to outsource manufacturing to countries like Vietnam and Bangladesh, companies will stop doing so. As AI and robotic technology continues to improve, more and more multinationals will replace their factories in the developing world with AI-powered facilities staffed with robots. And make no mistake: the intelligence and capacity of the ‘industrial AI’ in question is expanding at a rapid rate.
Nobody would argue that factory jobs or the like are perfect. But the alternative is surely worse. Take Nike’s factories in Indonesia, which employs over a hundred thousand people. These factories are not ideal workplaces by any stretch of the imagination — wages are low, hours are long, and health and safety is at perennial risk. But if Nike automates those factories and makes all its workers redundant, they won’t be the only one looking for another job — many thousands of others will be too, at the same time, in the same place.
Some pundits argue that the automation of entire industries like manufacturing, and the corresponding loss of jobs, will balance out with new jobs. The argument goes that just like other technological revolutions, AI will produce a new set of careers — AI ethics experts, suggests KPMG, or ‘AI architects’. I’m sure that on some level, this will happen (the tech giants alone could use some more ethics experts).
But you can imagine that these new high-tech careers would be of very little solace for the average factory worker. New AI jobs are overwhelmingly likely to go to educated, technologically skilled workers in the United States or other developed economies – not manufacturing workers in Bangladesh. The research available suggests that adjusting to a new career in the wake of automation requires resources, further education and often a welfare net: all favouring the unemployed in wealthier states, not developing economies.
This economic risk is compounded by the fact that developing countries are spectacularly ill-equipped to manage the health and welfare consequences of mass redundancies caused by automation. Assisting such a large number of displaced workers represents a gargantuan challenge for states already beset with structural problems like poverty and corruption.
What does this all mean? The technological wave of AI and robotics is likely to inflict the most damage on the exact countries that are least able to respond. At minimum, emerging economies face a torrid period of short and medium-term pain as they adjust to a post-AI economy. It is not even clear what this adjustment might look like — how can these countries improve their position in a world without jobs for their millions of low-wage workers?
Whether the eventual disruption is merely short-term devastation or a truly horrifying long-term crippling of emerging economies, there can be no doubt what will happen to global inequality. The technologically advanced nations (also being the world’s richest) will benefit economically at the same point that manufacturing-reliant countries suffer. Yes, there will be productivity gains, but exclusively in the developed world – economic growth in the United States will be of no help to an overwhelmed Thailand.
Chances are, this polarising shift will benefit me and you. Automation will make Amazon or H&M shopping sprees even more affordable (if not exactly advisable). We might even get bigger pay rises than we would otherwise get.
But as I’ve tried to lay out in this piece, the underlying technological shift causing these economic movements – AI and automation – carries a human cost. Worse, that human cost is uneven: most of the economic pain will fall on certain developing countries, making the damage more concentrated and harder to manage.
Despite the malign consequences, nothing can stop the pace of automation – the potential productivity gains make AI an irresistible option for profit-driven companies. This is particularly so given that the technology underpinning the automation will only get better.
The least we can do in the circumstances is acknowledge that lower prices for you and I represents a trade-off between capitalist incentives and the economic interests of developing countries. Real humans will suffer – AI isn’t (just) playing games anymore.