The onward march of cognitive capitalism

Until the Industrial Revolution upended the economic order, the world was run by mercantilists in fancy clothes. Wealth was seen as a scarce and finite resource. The rich and powerful squashed the poor. Subsistence farming generated no surplus consumer demand, meaning the global economy stagnated. This was a zero-sum world.

In a provocative new essay, the technology writer Venkatesh Rao asks whether we might be facing a similar situation today as the AI-driven Intelligence Revolution unfolds. Only these days, it is “cognitive mercantilists”, as he calls them, who are fixated on the idea that human intelligence needs to be defended from the excesses of AI, rather than set free by it.

Swap the word intelligence for wealth, Rao argues, and you are harking back to the conceptual universe of the early 18th century. Cognitive mercantilists view intelligence as a precious resource to be tightly managed by the credentialed classes of professionals, such as academics, consultants and lawyers. Countries and companies scour the world to ransack the best brains and hoard them like gold. Conversely, they dismiss the less intelligent “masses” as depreciating assets at risk of becoming superfluous humans.

But Rao argues that AI will challenge today’s cognitive mercantilists much as the spinning jennies and steam engines of the Industrial Revolution challenged the mercantilists of the past, enabling a new class of innovators to emerge.

As described in Joyce Appleby’s history of capitalism The Relentless Revolution, which inspired Rao’s essay, this was how positive-sum capitalism arose out of zero-sum mercantilism.

Today, billions of people can access increasingly powerful AI at negligible cost. Armed with these remarkable tools, anyone with wit will be able to pursue their own interests, find new ways of working, start their own microbusinesses and generate their own wealth. AI will spur the next “Industrious” Revolution. And AI agents, which will one day outnumber humans, will themselves provide an additional source of consumer demand.

“The cognitive mercantilist sees a world threatened by an oversupply of intelligence and an impending surplus of useless minds,” Rao writes. “The cognitive capitalist sees the possibility of an almost limitless expansion of demand, and a world full of minds whose potential has barely begun to be explored.”

This is a seductive theory. But it can be challenged on several grounds.

In the more conventional interpretation of events, AI is further concentrating wealth and power in the hands of corporate oligarchs. Some argue it may even serve as a tool of renewed enslavement rather than empowerment. The US stock market has never been as highly valued or as concentrated as it is today.

The data centres that are mushrooming across the world are seen as the “dark satanic mills” of our times, sparking a popular backlash against AI, especially in the US. According to a Pew Research Center survey, 40 per cent of Americans think AI will have a negative impact on society over the next 20 years with only 16 per cent expecting it to be positive. The technology is viewed as something to be resisted, not embraced.

That said, 49 per cent of Americans now use AI chatbots, one of the fastest uptakes of any technology in history. Some scientists also marvel at AI’s ability to open up new research pathways, such as AlphaFold, Google DeepMind’s AI model that has predicted the structures of more than 200mn proteins. And a fresh generation of AI-enabled start-ups, applying the technology in an ever-widening range of fields, is growing faster than previous cohorts of companies and at lower cost. The history of technology suggests that many of the most profitable opportunities lie at the frontier. “People can now use AI tools to increase the scope of what is possible and create more resources for themselves,” Alex Imas, director of AGI economics at Google DeepMind, tells me.

What is currently lacking, in the view of Imas and two co-authors of a recent paper on the economics of intelligence, are the organisational structures needed to harness the full capabilities of humans working with AI agents, which they call a “genuinely new kind of economic input”. Just as superintelligent companies emerged to co-ordinate the varied but complementary skills of humans, so we need new organisations to exploit joint AI and human intelligence. If this happens, that has big implications for the AI industry: how well AI agents collaborate with humans and each other will be just as important as what they can do.

john.thornhill@ft.com