Economist: Smart Machines
The Age of Smart Machines
Brain work may be going the way of manual work
IN HIS first novel, “Player Piano” (1952), Kurt Vonnegut foresaw that industry might one day resemble a “stupendous Rube Goldberg machine” (or as Brits would say, a Heath Robinson contraption). His story describes a dystopia in which machines have taken over brain work as well as manual work, and a giant computer, EPICAC XIV, makes all the decisions. A few managers and engineers are still employed to tend their new masters. But most people live in homesteads where they spend their time doing make-work jobs, watching television and “breeding like rabbits”.
It is impossible to read “Player Piano” today without wondering whether Vonnegut’s stupendous machine is being assembled before our eyes. Google has designed self-driving cars. America’s military-security complex has pioneered self-flying killing machines. Educational entrepreneurs are putting enlightenment online. Are we increasingly living in Vonnegut’s dystopia? Or are the techno-enthusiasts right to argue that life is about to get a lot better?
Two things are clear. The first is that smart machines are evolving at breakneck speed. Moore’s law—that the computing power available for a given price doubles about every 18 months—continues to apply. This power is leaping from desktops into people’s pockets. More than 1.1 billion people own smartphones and tablets. Manufacturers are putting smart sensors into all sorts of products. The second is that intelligent machines have reached a new social frontier: knowledge workers are now in the eye of the storm, much as stocking-weavers were in the days of Ned Ludd, the original Luddite. Bank clerks and travel agents have already been consigned to the dustbin by the thousand; teachers, researchers and writers are next. The question is whether the creation will be worth the destruction.
Two academics at MIT’s Sloan Business School, Erik Brynjolfsson and Andrew McAfee, have taken a surprisingly Vonnegutish view on this: surprising because management theorists like to be on the side of the winners and because MIT is one of the great strongholds of techno-Utopianism. In “Race Against the Machine”, their 2011 book, they predict that many knowledge workers are in for a hard time. There is a good chance that technology may destroy more jobs than it creates. There is an even greater chance that it will continue to widen inequalities. Technology is creating ever more markets in which innovators, investors and consumers—not workers—get the lion’s share of the gains. The Brynjolfsson-McAfee thesis explains one of the most puzzling aspects of the modern economy: why so much technological creativity can co-exist with stagnating wages and mass unemployment.
A new study by the McKinsey Global Institute (MGI), “Disruptive technologies: Advances that will transform life, business and the global economy”, shines a light on this problem and produces lots of examples of the way the internet is revolutionising knowledge work. Law firms are using computers to search through masses of legal briefs and precedents. Financial companies are using computers to monitor news feeds and make financial bets on the basis of the information they uncover. Hospitals are using robots to perform keyhole surgery.
The rate of progress, says MGI, is set to increase dramatically thanks to a combination of Moore’s law and the melding of three technologies: machine learning, voice recognition and nanotechnology. Tiny computers will be able to perform jobs once regarded as the peculiar preserve of humans: most middle-class people will soon have access to electronic personal assistants (to book flights or co-ordinate diaries) and wearable physicians (to keep a permanent watch on their vital organs).
MGI puts a typically positive spin on all this. It argues that being spared relatively undemanding tasks will free knowledge workers to deal with more complex ones, making them more productive. It argues that the latest wave of innovation will be good for both entrepreneurs and consumers. Small businesses will be able to act like giant ones, because cloud computing will give them access to huge processing power and storage, and because the internet destroys distance. Innovators will be able to test their new ideas with prototypes, then produce them for niche markets. Consumers captured much of the economic gains created by “general-purpose technologies” like steam and electric power, because they stimulated competition as well as increasing efficiency. MGI reckons that so far they have captured two-thirds of the gains from the internet.
Nevertheless, MGI’s study has some sympathy with Messrs Brynjolfsson and McAfee. It worries that modern technologies will widen inequality, increase social exclusion and provoke a backlash. It also speculates that public-sector institutions will be too clumsy to prepare people for this brave new world. Policymakers need to think as hard about managing the current wave of disruptive innovation as technologists are thinking about turbocharging it. For one thing, the purpose of education systems, and the skills and knowledge that they impart, will need to be rethought: “chalk and talk” instruction will be done best by machines, freeing teachers to become more like individual coaches to their pupils.
Knowledge-intensive industries will also have to rethink cherished practices. For a start, in an age in which information and processing power are ubiquitous, they will have to become less like guilds, whose reflexes are to regulate supply and restrict competition, and more like mass-market businesses, whose instinct is to maximise the customer base. Innovation will disrupt many areas of skilled work that have so far had it easy. But if we manage them well, smart machines will free us, not enslave us.