No seat belts in future smart cars?

Major automakers, for decades, heaped scorn on air bags and resisted making them standard equipment. At the start, 40 years ago or so, most people could not be persuaded to buckle up; this was before states passed laws mandating seatbelt use. Soon enough, interest turned to the air bag, a "passive" instrument that required no action by the driver and the passenger. If their car crashed, a balloon-like cushion would pop up on its own to protect them. Naturally, installing this device would make a car more expensive. In their fashion, the big auto companies balked, a stance that they maintained for many years, until their resistance finally wore out.

Not that air bags were perfect. They were initially designed with 165-pound adults in mind. That made smaller people, especially children, vulnerable to injury from the explosive force of the air bag's opening. In the 1990s, 175 people died in this manner, more than 100 of them children, according to the United States' National Highway Traffic Safety Administration. Terrible as that toll was, it lay in the balance against nearly 6,400 lives said to have been saved by air bags in that decade. Later improvements corrected many of the one-size-doesn't-fit-all shortcomings, though problems remained. A New York Times investigation, its results published on Friday, found that exploding air bags produced by a Japanese supplier were linked to two deaths and dozens of injuries in vehicles from Honda and other automakers. Slowly across the past decade, millions of potentially troubled vehicles have been recalled.

Will future vehicles even need these gadgets? Cars are becoming marvels of artificial intelligence, able to avert accidents entirely with sensors that, for instance, can warn a motorist that he or she is veering perilously into someone else's lane or getting too close to the vehicle up ahead.

The grail would be to remove the greatest threat of all to road safety: the driver. Human error is believed to be responsible for 90 percent of automobile crashes, making it the leading cause of the 33,000 motor vehicle deaths a year in the United States. Google and some major auto manufacturers are busily experimenting with autonomous cars, meaning driverless. Robots, the thinking goes, will sense lurking danger and take corrective action. A robot's judgment is not clouded the way a human's can be. It does not have one too many at a favourite bar. It does not nod off at the wheel. It does not succumb to road rage, at least not yet as best as anyone can tell.

But will humans happily yield control of their beloved cars to a robot? The immediate reality of artificial intelligence software at the wheel is hardly worry-free. If cars are effectively computers, could malicious hackers figure out how to seize control of them, even cause mass death? More prosaically, what if a computer on wheels freezes, as any computer is prone to do? Say it freezes for a mere three seconds. In three seconds, a car going 65 miles an hour will cover almost the length of a football field. Imagine travelling at that speed over that distance with no one - or no thing - effectively in control. Flaws that in the last few years forced companies like GM, Toyota and Honda to recall millions of vehicles are not likely to reassure the techno-wary.

Then there are privacy issues. Cars today can collect, in some instances for the benefit of auto insurance companies, all manner of data, like how fast people drive, how hard they brake, where they have travelled. Then again, auto companies were not alone years ago in chafing at innovations like seatbelts and air bags. Many motorists also had to be convinced. In time, they were. So maybe they will get used to being shepherded by robots.

©2014 New York Times