How low wages could keep productivity boosting technology at bay
By R.A. | LONDON
TECHNOLOGY companies from Google to Audi have made remarkable strides in autonomous vehicle technology over the past few years. This progress is the more impressive given the fact that a decade ago technologists considered driving to be a near un-automatable task. Despite the extraordinary pace of improvement, however, driverless cars still attract plenty of sceptics.
Some reckon that regulators and lawyers will keep the cars from reaching their full potential. They might do, though many local governments have been surprisingly open to crafting rules to accommodate the new driverless vehicles. Other critics argue that remaining technological hurdles could prove near-insurmountable. Yet such worries look overstated, both because recent progress suggests that obstacles are easy to overestimate and because remaining shortcomings ever more closely resemble those shared by humans.
Writing at Slate, for instance, Lee Gomes frets that driverless vehicles struggle in unfamiliar territory when they lack good maps, can make errors when sun blinds their cameras, and are occasionally caught out by the unexpected appearance of new traffic signals. Human drivers, of course, share these weaknesses, and others: like difficulty operating in adverse weather conditions. The big difference between driverless vehicles and humans, in these cases, is that the computer can be programmed to behave cautiously when stumped, while humans often plough ahead heedlessly. When critiquing driverless cars it is often useful to recall that human drivers kill and maim millions of people each year.
Ironically, the biggest obstacle to widespread use of driverless vehicles, over the next decade or two at any rate, may be the effects of rapid technological progress in other parts of the economy. As a recent special report explains, technological change over the last generation has wiped out many middle-skill jobs, pushing millions of workers into competition for low-wage work. That glut has contributed to stagnant wages for most workers, and low pay has in turn reduced the incentive to firms to deploy labour-saving technology. Why automate, when there is an enormous stock of cheap labour available? At the same time firms like Uber are making the use of hired cars cheaper and more convenient, reducing the attraction to many households of owning and driving their own personal vehicles.
The combination of Uber and cheap labour could pose a formidable threat to the driverless car. The cost of the sensors and processors needed to pilot an autonomous vehicle is falling and is likely to fall much more as production ramps up. Yet the technology is still pricey, especially compared with a human, which, after all, is a rather efficient package of sensory and information-processing equipment. At low wages, a smartphone-enabled human driver is formidable competition for a driverless vehicle.
It would be a remarkable irony if the driverless car—in many ways the symbol of the technological revolution that is now reshaping modern economies—fails to materialise as an economic reality thanks to the disemploying power of other technologies. It would also be a perfect illustration of how stagnant wage growth can cause slow growth in measured productivity.
The possibility of a world in which a rather large share of the population works as drivers, simply because human labour has gotten too cheap to automate out of the job, should focus minds on the nature of the policy challenge economies are beginning to face. Is work—and the link between work and the earning of an income sufficient to live on—so important to society that we should want millions of people to function as meatware: doing jobs sensors and computers could and would do if only there were not an excess supply of humans needing to work in order to afford food and shelter?
That's not a rhetorical question. It's a genuine puzzle that societies will find themselves confronting in coming decades. It will be obvious to many people that the answer is no and just as obvious to many others that the answer is yes. I cannot begin to say which side will win the argument.
