Inside the Information Society: The second machine age - reasons to be fearful?

Each week David Souter comments on an important issue for APC members and others concerned about the Information Society. This week's blog post looks at what's meant by the 'second machine age' - and how it should reshape our thinking.

When I was a child (in the 1960s), there was a popular TV show in Britain called Tomorrow's World. Each week it showed the wonders new technology had in store for us – like ATMs and pocket calculators.

Much of what's written today about the ICT revolution reminds me of that programme: it's about the gadgets, devices, applications we can use to make some aspects of our lives a little easier. But it's important, too, that we look at the bigger picture - at the aggregate changes that are happening as a result of changes in IT.

It's these big picture changes that will make up the Information Society, and they're very different from those that we imagined at the World Summit a decade and more ago. Those of us who are concerned with development, or rights, or governance need to pay more attention to this.

There's a growing body of literature on the bigger picture: some positive, some negative; with reasons to be fearful alongside reasons to be cheerful. This week a summary and some thoughts on one of the most influential books about it, Erik Brynjolfsson and Andrew McAfee's The Second Machine Age, and Brynjolfsson's latest update.

What is the second machine age?

Here's their basic line of argument:

For most of human history, they say, we relied on our own physical and mental resources. Two periods of innovation – two 'machine ages' – have changed this.

The first machine age was the Industrial Revolution of steam engines and, later, electricity. The machines that we developed then increased the physical power at our disposal. This led to huge improvements in productivity, and mass production. It's enabled sustained growth over two centuries in the quantity and quality of goods and services, in human prosperity and in the population that the planet can support.

We're now, they say, in the early stages of a second machine age, this time one of digitalisation and information technology. The machines we're developing today increase the mental power we have at our disposal: they "do for brains what the Industrial Revolution did for muscles."

This second machine age is driven by three characteristics of ICT development – that it's exponential (not just growing in scope and scale each year, but actually accelerating), digital (enabling perfect replicas to be made at near-to-zero cost) and "combinatorial" (drawing together things that once were separate). (Others, confusingly, have referred to this "second machine age" as "the fourth industrial revolution".)

The result, they say, is "bounty". By which they mean not just improvements in specific goods and services, or incremental changes to services like health or education, but fundamental changes in economic, social and political relationships; changes that will be as profound as those of the Industrial Revolution and more.

What next? A second wave

Brynjolfsson updated the argument in a lecture earlier this month. A new book is also on the way. We're now, he says, entering the second wave of the second machine age. Here's what he meant.

In the first wave of the second machine age, he says, we taught machines what we know. That gave us many new resources; enabled us to do more complex things more simply and more cheaply. But, like those of the Industrial Revolution, our new digital tools complemented our abilities, rather than substituting for them.

The second wave's driven by machines that are learning on their own. We give them data, and they teach themselves how best to use them. In the last few years, and much faster than expected, computers have become much better at recognising patterns, developing strategies and solving problems – and this progress is accelerating. He cites three areas where, as a result, machines are now substituting for us not just complementing us:

  • interaction with the physical world – for example autonomous vehicles, high-speed factory production and facial recognition;
  • language – for example, speech recognition, near-simultaneous translation and "creating narratives" (like computer-generated news feeds on sports results or business data);
  • problem-solving – for example, game-playing, analysis of complex data-sets like legal precedents, and the use of algorithms to select employees.

And the impact?

The Information Society that results won't look much like that which we envisaged at the World Summit in 2005. Or that discussed in most literature about ICT4D or ICTs and human rights. It will be far more transformative and far harder for us to control.

Brynjolfsson and his colleagues are positive about the potential of their second wave. They see much greater "bounty" resulting from it, as it improves efficiency, reduces costs, enables innovations that can/will bring joy, sustain health and improve welfare. But they recognise, as well, the challenges of rapid, radical transformation. Two in particular.

First, there's what they call "the great decoupling". In the first machine age, employment and pay levels rose alongside productivity. That's no longer happening. Robots and computers are already displacing routine manufacturing and middle-level jobs. Second wave developments like autonomous vehicles and algorithmic decision-making will displace people in more job categories: not just drivers but also professionals (such as lawyers). As a result, new technology is increasing economic inequality, and that's likely to continue.

They're not as pessimistic here as some. Important types of work, they think, will not be displaced for quite some time – some physical jobs, creative work, jobs requiring empathy or entrepreneurial skills. But they're sure that we'll need radically new approaches to education for the Information Society (not just tinkering with skills), and may need to think of new ways of distributing income than employment (universal basic income, anyone?).

Second, Brynjolfsson and his colleagues argue that we're badly prepared. The second machine age is coming fast and is poorly understood by governments and other stakeholders. Business-as-usual won't deal with it, so "we need to reinvent our businesses, economy and society to keep up with accelerating technology."

So what is to be done?

I've summarised at length because this matters. We all – in governments, business and civil society – think a lot about what we'd like to do with ICTs. We need to think much more about where ICTs are leading us, how we must adapt to this new Information Society, and whether we'll have any chance to shape it.

First, employment. Predictions that the second machine age will displace jobs and polarize employment are increasingly widespread. Brynjolfsson's "less vulnerable" types of work don't make up most of today's job markets. Even small reductions in demand can have severe impacts on individuals' job prospects and on local labour markets. Reskilling won't be enough to give everyone a job. Wage stagnation and growing inequality pose risks to social and political stability.

Second, governance. I'll point to two phenomena. One's the growing power which new technology gives those who finance and deploy innovation: increasingly, today, international corporations. The other's that increasingly sophisticated technology is not being matched by sophisticated governance. On the contrary, the Information Society's emerging in a world whose politics is retreating from globalisation; increasingly populist, nationalist and authoritarian.

In conclusion

Much of our thinking, and most of our debates, about ICTs, development and rights are concerned with how we'd like ICTs to change the past, rather than how ICTs will change the future.

In 2003, at the World Summit on the Information Society, the international community committed itself to build "a people-centred, inclusive and development-oriented Information Society." But the Information Society we then thought possible is entirely different from the Information Society that's now emerging. How "people-centred" is robotics?

In 2012, the UN Human Rights Council declared that "the same rights that people have offline must also be protected online." But the context in which those rights exist, and the meaning of "online", are changing beyond recognition. How 'empowering" are decisions made for us by algorithms?

In 2014, the international community adopted the 2030 Agenda for Sustainable Development, a core aim of which is that "no-one should be left behind." And here's Brynjolfsson – a digital optimist, remember: "Digital technologies change rapidly but organisations and skills aren't keeping pace. As a result, millions of people are being left behind."

We've built an enormous literature about the Information Society that we'd like. It's time we refocused on the Information Society that we're going to get – one which is designed by technologists, not social visionaries or politicians; delivered by businesses, not international agreements or governments; which is much, much wider than the Internet and much more challenging than "enhanced cooperation" and "net neutrality".

Next week I'll look back at one of the classic texts of ICT4D – and asked what's really happened to the price of fish.

Image: Erik Brynjolfsson at MIT Sloan CIO Symposium 2013. By Gordon Haff used under Creative Commons license.

More about David Souter