Autonomous Vehicles: What’s Causing the Delays, and How They Can Get Off the Dime

Per recent posts, the past year has been rather disheartening for those of us who think we need progress on self-driving cars.  What has been the problem, and how can they get moving again?

The best article describing the state of that field so far came out in the July 17^th New York Times.  Neal E. Boudette’s “Despite High Hopes, Self-Driving Cars are ‘Way in the Future’,” started by saying that “a year ago, Detroit and Silicon Valley had visions of putting thousands of self-driving taxis on the road in 2019, ushering in an age of driverless cars,” which hasn’t and won’t happen.  The piece’s sources blamed the lack of progress on being able to deal with “all kinds of crazy things on the road,” and Boudette also named small related news items, recapped the fatal accident, quoted evidence of Elon Musk’s delusional hubris, and then got the article to justify itself by naming the need for “micro maneuvers” such as understanding other drivers are looking for a parking space so should not be followed closely and saying that “the technology is available now to create a car that won’t hit anything,” even if it would “constantly slam on the brakes.” 

From what I have read and not read, I see five reasons why progress has almost ground to a halt.  First, not only overreaction to that single death, but expected overreaction, as the firms seemed to pull back soon thereafter without receiving much actual pushback.  Second, companies’ testing has, thus far, not emphasized creating algorithms mimicking the thoughts of actual drivers.  Third, as I have read recently about artificial intelligence, massive efforts such as this often go through slow stretches in research intensity, which author and professor Nick Bostrom called a “winter,” or “period of retrenchment” – we clearly have another one here.  Fourth, the legal and regulatory climate, despite the July 31^st Yahoo Finance report that “U.S. Congress seeks to jump start (a) stalled self-driving car bill” to allow them more, has been intimidating if not actively discouraging, perhaps to the point where companies have focused excessively on being stopped.  Fifth, there has been insufficient emphasis on implementation – I saw that when working with information technology technicians more comfortable keying on the clean and promising future than on dealing with the grit of making things actually work, and recognize, or at least strongly suspect, it again. 

What, then, are the solutions?  Here are six.  One, researchers need to catalog what Boudette called “corner cases,” where people disobey traffic laws, and concentrate on solving them.  Two, it is time for them to quantify how drivers actually think when faced with these problems.  In Kurt Vonnegut’s Player Piano, those making automated barbers copied the exact movements of people doing that job – let’s do that here.  Three, they should implement Boudette’s “car that won’t hit anything” and see where it could be used – the technology could progress from there.  Four, we need more driverless shuttle buses in limited, well-defined settings, which is about the lowest-hanging fruit actually constructive.  Five, manufacturers and others should push for more freedom and more places to use these vehicles, which may, given the recent Congress event, be easier than they think. 

Sixth, and finally, those on the side of this technology need to do interviews, give presentations, write articles, issue news releases, do radio spots, appear on TV shows, and put in Internet advertising, all emphasizing the potential for slashing the 30,000 annual driver-caused American deaths along with other autonomous-vehicle advantages and showing how close we really would be if we can tolerate a few more accidents.  There were times when it was more acceptable for efforts thought of as American projects to cause tragedies along the way.  Eighteen people have died during space flights and 96 perished during Hoover Dam’s construction, not to mention such numbers as over 400,000 United States soldiers killed in World War II.  It is time for us to consider driverless vehicle implementation necessary for the country, and give it the same status.  Then, as we know about our countrymen and from our history, it will succeed.