The algorithm society and its discontents

In my view, the most profound and insightful work of political economy written in the 2010s was neither a journal article nor a monograph nor a book in the traditional sense. Rather, it was an online symposium. In Red Plenty: A Crooked Timber Book Event, scholars and intellectuals, convened by political scientist Henry Farrell, used a new mode of print-communication to react to Francis Spufford's very interesting book Red Plenty.

Spufford had analysed the Soviet Union's stunningly unsuccessful attempt to use bureaucracy and mathematics to build a better society than could be achieved using markets. Yet every time I return to Red Plenty: A Crooked Timber Book Event, I am struck by its contributors' insights into the insurmountable dilemmas generated by the modern market economy itself. I am also still struck by how successful the "book event" was in using new technologies to drive a qualitative shift in how we communicate and come to understand the world together.

I have been thinking about these issues because Farrell recently published a new article, "The Moral Economy of High-Tech Modernism." He and the sociologist Marion Fourcade argue that the internet and its progeny (what they call "high-tech modernism") are changing the world in ways that are as profound as the rise of the market economy and the bureaucratisation of society under the modern state.

This argument concerns the very foundation of human society. As individuals, we humans are each weak and helpless. Only with the knowledge that we gain from life within a culture are we able to survive. But to have a culture or produce anything, we need to organise and coordinate ourselves in a collective intelligence and distributed entity. For thousands of years, we generally have had three different modes of maintaining such organisation.

The first is redistribution: information, resources, and useful products flow into a centralised pool and then flow out again in the form of assigned tasks, tools, rewards, or social support. The second mode has been reciprocity: each household is linked to a few others in long-term gift-exchange relationships that tend toward a rough balance. And, because there are only six degrees of separation in most human societies, one unit's needs will affect the actions of many others. Lastly, there is democracy: people use debate and discussion to reach a rough consensus and achieve broad-based support for an agreed-upon plan.

Of course, each mode of organisation entails a mode of distribution and authorisation, to answer the question of who should get more of the good things (who should be "more equal" than the others). With redistribution, power accrues to the one at the center of the system; with reciprocity, to those with the most resources and friends; and with democracy, to those with the silver tongues.

With the coming of modernity, we added two more modes: the market economy as engineered by the business class; and bureaucracy, as engineered by the modern state. The market is unmatched as a tool for crowdsourcing solutions to problems. But, in practice, its scope is limited to satisfying the demands of the rich, by ensuring the efficient use of those things that just so happen to command a market price.

Similarly, bureaucracy is uniquely powerful and capable in its ability to classify and standardise things, which allows it to see the bigger picture in ways that a reciprocity or redistribution system cannot. But, of course, it also can give rise to many inefficiencies.

Our mighty, but deeply flawed and unequal civilisation was built by adding markets and bureaucracy to our three original modes of organisation. But now, Farrell and Fourcade warn that we are adding a sixth mode: the algorithm.

According to techno-optimists, a society of algorithms would be much better than anything we can hope to create with markets and bureaucracy. Unlike a market, an algorithm is not restricted to seeing only the money demands of the rich and the money costs imposed by those who have managed to claim property rights. And unlike a bureaucracy, an algorithmic society will not force you, a square peg, into a round hole.

No longer will "experts" decide what category you should fall into. Instead, affinity groups will spring up spontaneously from the revealed preferences expressed by people's words and actions. Resources will be mobilised to serve each individual by tapping into the unique power of economies of scale.

Is this a hope that we should all share? To be sure, when bureaucracy arrived as a new mode of organisation, it erased tacit forms of knowledge, disrupted the messiness of people's lives, and forced people into categories that were most useful to those holding the levers of power. Equally, markets introduced massive, costly new externalities by prioritising the needs of the rich. But is there any good reason to think that an algorithmic society would fix these flaws, or that it would not introduce new, massive problems of its own?

The economics Nobel laureate Daniel Kahneman's concept of thinking "fast" and "slow" can help us see what is at stake, here. An algorithmic society would serve only our "thinking fast" side, constantly seeking "engagement" – meaning fear and rage. Just as markets cater to the rich, algorithms cater to our worst impulses.

A clickbait society is no one's vision of Utopia. Yet that, I fear, is where current trends are carrying us.

J Bradford DeLong, Professor of Economics at the University of California, Berkeley, is a research associate at the National Bureau of Economic Research. He was Deputy Assistant US Treasury Secretary during the Clinton Administration, where he was heavily involved in budget and trade negotiations. His role in designing the bailout of Mexico during the 1994 peso crisis placed him at the forefront of Latin America's transformation into a region of open economies, and cemented his stature as a leading voice in economic-policy debates.

Disclaimer: This article first appeared on Project Syndicate. It has been edited and published by a special syndication arrangement.