John Thornhill
The commander of D-Day knew a thing or two about planning. After plotting the 1944 assault on Nazi-occupied France, Dwight Eisenhower would say: “Plans are useless, planning is essential.”
Eisenhower knew that in the heat of battle no plan would survive the first shot. But that did not mean that planning was a waste of time. If you failed to prepare then you were preparing to fail.
Eisenhower’s dictum is worth thinking about as we consider the possibility of mass technological unemployment. We have no idea how the future will pan out and therefore we can devise no sensible plan. But, given what we already know about the rise of the robots, it would be reckless not to plan for the possibility of extraordinary disruption.
Of course, there are those who assert that mass technological unemployment will never happen. Such fears are the latest manifestation of the Luddite fallacy. The future will turn out just like the past: we can easily see the jobs that will be destroyed by technological progress but can never foresee the ones that will be created. The latest employment report from Cognizant shows that one of the jobs most in demand is cyber calamity forecaster. Such a role could not have been imagined 25 years ago. If work can be defined as unsolved problems, then we will always have plenty of those.
But whereas economists think in terms of curves, technologists live in a world of step changes. We should, at least, admit the possibility that we are facing an unprecedented step change in technology as cognitive computing replicates ever more human capabilities.
For the past few years, I have attended discussions at the Economic Singularity Club, a loose group of technologists, academics and writers who think the threat of mass technological unemployment is worth taking seriously. Last week, the club published a book of short stories by some of its members (including me) speculating on what the world might look like in 2045.
The good news is that the positive stories outnumber the negative by two to one. Many of the human contributors thought we would be able to adapt successfully. In a further boost to team human, those stories written by AI programs, trained on 300,000 stories submitted to Reddit, proved reassuringly patchy.
Calum Chace, the club’s co-founder who has written extensively on the disruptive threat of AI, argues there is a danger of succumbing to the Reverse Luddite fallacy. Just because mass technological unemployment has not happened before, that does not mean it will never happen.
“We have been told that we are at the beginning of the Fourth Industrial Revolution. But that is dangerously misleading,” he says. “We are at the early stages of an information revolution that will have a much bigger impact than all previous revolutions.”
Recommended
It is sometimes said that technological upheavals — like Ernest Hemingway’s description of bankruptcy — happen gradually, then suddenly. To appreciate just how fast technology is evolving it is worth listening to Chris Bishop, who runs Microsoft Research in Cambridge.
He argues that software may see startling improvements over the next few years, as machine learning becomes pervasive. “I suspect that we may be seeing a singular moment in the development of software,” he says, speculating that Moore’s Law, which predicted an exponential increase in computing power, may now, in effect, be flipping from hardware to software.
Carefully managed, this development could yield phenomenal benefits in the field of healthcare, for example. The techno-optimists argue it could even lead to a world of radical abundance or fully automated luxury communism. That explains the relative optimism of some contributors to Stories from 2045.
But it is also possible that the next wave of technological change might simply exacerbate adverse trends: the concentration of corporate power and acute divisions in society. That suggests we may need more imaginative solutions than our existing political institutions can devise. The British parliament already appears to have blown a fuse in trying to handle a relatively straightforward problem such as Brexit.
The stock responses to the technological challenge, aired at countless conferences on the future of work, are to retrain workers, redignify professions that can only be performed by humans (such as nursing and caring), and redistribute wealth. But a more radical reshaping of the way we run our societies may yet be required. Perhaps, to warp a concept, we may need a moment of political singularity. Better start planning now.