Friday, November 20, 2015

Clearing the Exhaust Fumes about Self-Driving Cars and Jobs

Two large articles and one small one about a new technology area hit the press last week.  The first, in The Economist, “If Autonomous Vehicles Rule the World from Horseless to Driverless,” reminded me of a landmark February 2011 Economist Technology article describing how 3D printing “may have as profound an impact on the world as the coming of the factory did” and “is likely to disrupt every field it touches.”  It discussed the current and emerging state of self-driving technology and made predictions, some framed as facts in the future tense, that such autos “will also challenge the very notion of car ownership,” that “coming generations will consider the era of car ownership to have been much stranger,” and even that the thought of automotive executives that many will continue owning vehicles “sounds a lot like wishful thinking, if not denial.”  The second, “The Dream Life of Driverless Cars” by Geoff Manaugh, printed November 15th in the New York Times Magazine, presented a projection that automated vehicles would make up “up to” 75% of those in 2040, but was less giddy about their future, naming dependence on “cars being able to navigate the built environment,” incomplete areas of their software’s development, and the inherent difficulty of interpreting what their programming needed to drive flawlessly while ignoring that, such as “glares, reflections, and misunderstood signs” which it would not.  The third article was a simple news item in The New York Times and picked up elsewhere, reporting that on November 12th a driverless car in Google’s hometown of Mountain View, California had been pulled over for obstructing traffic by going 24 miles per hour in a 35 zone.  The police did not issue a citation, but indirectly pointed out a flaw in its software, by directing its two non-driving operators to briefly take it off the road if other vehicles stacked up behind it. 

As I documented in an April post, 3D printing has not yet approached projections made almost five years ago, but will driverless vehicles justify the Economist clipping by 2020?  I offer eight observations.

First, there is no assurance whatever that automated cars would cut the number of those on the road by anything like the 80% to 90% one source claimed.  Simply having access to less expensive taxis, which is hardly a short-term certainty itself, would not cause the bulk of car owners to turn them in, as they live, work, and run errands in areas where they would need to wait too long for one, driverless or otherwise, to arrive.  Also, if there were fewer of them and their utilization were high, it would take even more time.

Second, if existing road and surrounding structures need to be standardized or otherwise modified to accommodate driverless cars, it may take decades.  The United States government and its constituents have done nothing comprehensive on national infrastructure needs as it is.  Accordingly, it may well prove to be another case where America is behind the world leaders, to the point where automated vehicles have widespread use elsewhere but not here.

Third, one thing authorities will need to address is preventing driverless cars from being hacked.  A one-ton-plus machine capable of moving at dozens of feet per second is a deadly weapon.  Even if malevolent people can succeed only in deactivating them, that could cause catastrophic transportation-system delays, especially if large numbers could be put out of commission at once.

Fourth, while self-driving vehicles seem to be safe so far – in fact, Google claims those it has made have caused no injuries or even at-fault accidents in 1.8 million miles – it is almost certain that bizarre circumstances, and programming bugs, will cause some spectacular ones sooner or later.  As a robotics professor put it in the Times Magazine story, since “we will have millions of these cars, the very unlikely will happen all the time.”  It is not inconceivable, for example, that a driverless tractor-trailer could interpret a stopped rainbow-colored van as an atmospheric phenomenon, as its software might not know that vehicles could have that color pattern, and run it down without even slowing.

Fifth, however, if we look at the reasons I cited for 3D printing not becoming massive, none apply to self-driving autos.  There can be no shortage of “killer aps” for these new cars, and one clear to materialize is their usability by those who should not or cannot drive otherwise.  As their new technology is easily duplicable software instead of physical parts, the automated ones almost guarantee to be cost-effective.   I also perceive no concern with either quality or performance, both of which have slowed the use of 3D technology, for self-driving autos as compared with ordinary vehicles.

Sixth, the safety of driverless cars will provide large incentive for their implementation and purchase.  With 32,719 Americans killed in motor vehicle accidents during the most recent data year of 2013, and 94% of those mishaps from driver error, that provides over 30,000 deaths with at least the potential to be eliminated.  Even if only 15,000 lives were saved, that would be almost 1½ times the number I projected, in a post last month, from bipartisan gun-law reform.

Seventh, as opposed to the ending of the Economist piece, automated vehicles would not mean an end to personal freedom.  The only difference would be, due to the likely requirement that such cars be fit with airliner-style black boxes, that information for all trips would be a matter of record.  With the share of Americans owning GPS-monitoring smartphones at 64% and rising, that seems likely to bother few.

Eighth, the number of jobs that self-driving cars will create or destroy will depend heavily on its acceptance.  Even though the technology stands to save many lives, the peculiar disasters that will happen along the way will unnerve many.  After the 1986 Challenger space shuttle disaster, a remarkable number of Americans thought space travel and exploration should be discontinued – and indeed the next launch did not take place until 31 months afterwards.  Others may, for whatever reason, be resistant to getting in driverless taxis or sharing the road with automated 18-wheelers. 

How will this new capability affect jobs?  We could be looking at a wholesale drop in the number of cabdrivers by 2020, or their numbers could still be strong a decade or more later.  Truckers, likewise, may keep their jobs, or see them severely decimated.  Overall, self-driving technology may be almost here, but we cannot presume to know its effect.  As with other jobs-related areas, I will keep track of developments and let you know.          


  1. This comment has been removed by the author.

    1. Yes! The article was connected with the Economist, and was at . One you certainly should see if you are doing research on the issue. Thank you for your interest!