The Post: Taking control of technology’s future direction

Read the original article in the Post

As we reach the end of yet another year that has not, perhaps, been the best the world has ever seen, it would be easy to feel defeated about the future. War, drought, famine, climate change: no end of things trouble us.

Technology has become another potential source of pessimism. Many people are understandably worried about the way that AI, in particular, seems to be evolving beyond our control, its economic threats at least as large as its promises.

But we should not be defeatist. That, at least, is the message of the best book I read this year, Power and Progress: Our Thousand-Year Struggle over Technology and Prosperity, by the Nobel economic laureates Daron Acemoglu and Simon Johnson.

Their basic contention is that technological development, AI included, can yield extraordinary economic benefits to humankind – but only if we point it in a specific direction. And we hold that power in our hands.

What Acemoglu and Johnson have in their sights is a naïve – and dangerous – version of the “productivity bandwagon” theory which, as espoused by the tech lords of Silicon Valley, holds that technological development is uniformly positive because its economic benefits automatically filter out across society. Citizens and their governments should, in this view, just get out of the way of technological change, and watch productivity rise and living standards soar.

The beauty of Power and Progress is that it delves deep into 1000 years of history in order to point out that this theory is not, in fact, universally true. I already knew that during the Industrial Revolution, many workers were temporarily worse off after being displaced by innovations like spinning machines. But even I was surprised to learn that the living standards of textile workers went backwards for an entire century.

Even more startlingly, it turns out that the mediaeval period had a fecundity of technological innovation far beyond what I had believed – including greatly enhanced ploughs, the first fireplaces and chimneys, and the development of mechanical clocks and eyeglasses – but the living standards of peasants stalled for 200 years.

Why is this? Acemoglu and Johnson’s great insight is that the “productivity bandwagon” operates only if two conditions are met.

First, a given innovation has to raise something known as worker marginal productivity – the additional contribution that each extra worker makes, for instance by increasing throughput in a factory. Innovation, in other words, has to make each extra worker more productive.

This has happened often throughout history. New forms of software, for instance, have helped car mechanics work more rapidly and productively.

Technological development can also create whole new tasks. Henry Ford’s assembly-line method for making cars may have rendered some workers obsolete, but it created a vast swathe of design, technical, machine-operation and clerical tasks. It also spurred the development of other industries – in oil, steel and chemicals, for instance – and new forms of employment, from inter-city freight through to the humble drive-in cinema.

In contrast, pure automation, the replacement of people with industrial robots, leads to fewer jobs and can effectively make workers less productive. Acemoglu and Johnson are also scathing about what they label “so-so automation”. Self-checkout systems in supermarkets don’t meaningfully increase productivity or bring whole new industries into being: they just shift the work of scanning items from employees to customers.

The second thing needed to make the productivity bandwagon a reality is an institution – a law, an organisation, a body of some kind – that ensures the gains from productivity are broadly shared with workers, in the form of higher wages. Powerful trade unions and minimum-wage laws are two such institutions.

Lacking collective power, mediaeval peasants and early Industrial Revolution workers unsurprisingly saw their living standards degrade. In the latter case, it was only once trade unions hove into view, and labour laws were reformed, that industrial innovation began lifting ordinary people’s wages.

The underlying point here is that we, as citizens, retain that most precious of commodities: choice. We do not have to submissively accept the technological changes espoused by our would-be overlords in Silicon Valley.

There are many ways we can ensure that digital technologies, including AI, bring widespread economic benefits, making workers more productive in their current jobs or creating new tasks for them to do. Government can, and should, provide subsidies for socially beneficial technologies, regulate surveillance technologies, protect people’s privacy and data ownership, and invest massively in lifelong vocational education.

These are not pie-in-the-sky ravings; Acemoglu, in particular, is probably the world’s most influential and broadly respected economist.

He also argues, correctly, that in political mobilisations like America’s late-19th-century Progressive movement, we see examples of ordinary people succeeding – against great odds – to overcome the domineering power of big capital and vested interests.

“The future path of technology”, Acemoglu and Johnson conclude, “remains to be written.”