TL;DR: it's up to us
This video, Will AI destroy the economy? from Gary's Economics is a must watch. I viewed it on 1.5x speed and it's a really useful, accessible message.
I asked my Little Robot Friend to create an overview with timestamps for those who aren't so fond of watching videos:
AI is raising familiar questions about wages and inequality. Some outlets, including Fox Business and the International Monetary Fund, argue that AI could narrow wage gaps and lift pay for low‑wage workers, on the basis that it can raise their productivity and bargaining position (00:00:09). Many economists hold a general view that better technology should push wages up, since productivity growth is usually seen as the main driver of wage growth over time (00:04:09). At the same time, there is a clash of narratives: one line of argument sees AI as a force that will displace workers and cut wages by increasing unemployment, while another sees it as a tool that will enhance workers’ productivity and make them more valuable to employers (00:08:56).
The industrial revolution is often used as a historical benchmark for thinking about AI. In Britain, new manufacturing methods and factory systems drove huge productivity gains, reorganising production around machinery and centralised workplaces (00:10:26). That transformation displaced skilled artisans, who saw their livelihoods eroded and, in many cases, their wages fall, as illustrated by the Luddites’ resistance and machine‑breaking campaigns (00:13:37). This period followed enclosure laws that pushed many farmers off common land and into towns, where they formed a surplus labour pool that factory owners could employ on very low wages in harsh and unsafe conditions (00:17:06). The outcome was a long stretch in which factories replaced artisanal work, but poverty and poor living conditions dominated, and significant improvements in everyday life for workers did not really appear until after the Second World War (00:20:42).
During that era, wealth distribution was highly skewed. The British elite accumulated vast fortunes from industrialisation, while workers were underpaid and had to organise and struggle for better pay and conditions through unions and labour movements (00:26:15). Rising living standards were driven to a large extent by these movements and the policies that redistributed resources, not simply by the technologies themselves (00:28:09). It is misleading to assume that AI on its own will improve living standards if similar political and social efforts are absent.
The familiar story that productivity growth automatically lifts wages depends on strong demand for goods and services. That assumption is under strain today, given concerns that the global economy may lack a broad, affluent customer base able to absorb ever‑increasing output (00:32:27). Earlier industrial revolutions, including in Britain and later in China, had access to expanding markets and relatively wealthy consumers to sell to, which helped sustain growth (00:32:35). If AI significantly boosts productivity in the current context, there is a real question about whether demand will keep pace, and one response might be to use technological gains to shorten working hours rather than keep pushing output higher (00:34:51).
John Maynard Keynes famously imagined a future in which people would work fewer hours as technology advanced, leaving more time for leisure and non‑market activities (00:38:37). That prediction has not really come to pass, largely because inequality has intensified, so the benefits of new technologies are not broadly shared. In such an unequal setting, paid work becomes central to gaining access to housing, healthcare, education, and to servicing debt that is often held by a small, wealthy group (00:41:55).
Ownership of technology sits at the heart of this discussion. Those who own AI systems and related infrastructure effectively control access to key resources and shape how gains are distributed (00:44:07). The experience of the industrial revolution shows that new technology can inflict serious harm on large numbers of workers, while enriching a small elite, and that broad improvements in living standards came only after sustained organising and redistributive policies (00:44:36). For AI, this suggests that what matters is not just the tools themselves, but who owns them and how the rules around them are set.
This leads to a focus on equity and collective action. Ordinary people’s ability to claim a fair share of the benefits from technological change, and to insist on decent wages, universal healthcare, education, and secure housing, is central to building a more equal economy (00:45:21). Without that, AI risks replaying familiar patterns of concentrated wealth and widespread precarity rather than delivering the broadly shared gains that headline narratives often promise.
What is the main difference, in your view, between the optimistic “AI raises wages through productivity” story and the historical account of the industrial revolution you have summarised here?