When discussing the future of
autonomous driving, the common refrain amongst interest groups automakers and
the public are the superior safety conditions these vehicles will produce.
After all, more than 37,000 people die in the US every year from car accidents
and 2.4 million are injured or disabled. Globally those numbers jump to 1.3
million and 50 million respectively. From a financial perspective, vehicular
accidents cost more than $231 billion annually in the US alone. And with
greater distractions from smartphones and unceasing levels of traffic
congestion, it’s no wonder the world is clamoring for safer and more autonomous
modes of transportation that remove human error from the equation. But in a
grand irony, the very safety standards the development of self-driving cars
promise are also hindering its advancement. The assumed infallibility of
autonomous cars will lead to misinformation and distrust among consumers
unwilling to relinquish control to a technology that is not guaranteed.
Though the term ‘safety’ is batted around among interested parties, stopping to produce a consensus on the actual standards of the idea produces a cacophony of discordant voices unable to reach agreement. Does autonomous driving need to produce fewer accidents than humans or none at all? The latter is impossible to achieve as no machine operates at 100% efficiency. And who determines the safety standards? Currently, US regulators and Silicon Valley have shifted the responsibility for determining these benchmarks as neither want to be liable for the inevitable botches in the system. Based on our experiences with computers and smart phones, technology can often fail and this is problematic for achieving massive adoption of autonomous driving. Will consumers be willing to relinquish control to a fallible robot, especially if liability and fault obligations are not clearly defined? You’d be hard pressed to find insurers willing to assume the financial and criminal liability in car accidents for technological errors. There should be legal definitions for the safety and responsibility of automated vehicles, but federal regulators are ill-equipped to address those issues.
Image Credit: Smart2Zero
While there might be a general consensus self-driving cars are better at driving than the population at large, human psychology dictates that an individual will assume he/she possesses better judgement than a machine. Based on the theory of illusory superiority, we overestimate our own abilities and underestimate the larger group. Why else do 93% of US motorists think of themselves as better drivers than the rest of the population. When autonomous cars inevitably have accidents, instinctively people will assume they could have done something differently to prevent the accident had they been in charge. Statistics will not support them, but it is difficult to convince a person of assumed inferiority as humans naturally assume they can avoid situations that befell others.
Furthermore, every automaker determines the level of “automation” of its vehicles, ranging from assisted parking and breaking to full automation, thus requiring different levels of attention and participation from its users. Will there then be levels of safety ascribed to various vehicles depending upon their abilities? Consumers will have to determine if it is worth assuming the increased expense of these vehicles if their attention is still required like traditional motor vehicle operation. Some insurance groups believe training (or an additional drivers test) will be required for owner/operators to certify they are aware of the specific capabilities and limitations of their automated vehicles. Allowing drivers on the road who assume their car possesses certain automated faculties has led to the industry’s most high profile accidents (i.e. 2016 fatal crash of Tesla Driving Model). Unfortunately, these added safety measures are likely to deter the purchase of self-driving vehicles if consumers are convinced the added time, expense and responsibility is more burdensome than traditional car ownership.
Until these gaps in understanding of promised safety standards versus reality are addressed, there will be massive stagnation of the adoption of autonomous driving. The largest motive to embrace the technology is security; if assurances are removed from the equation, you remove demand. Automakers banking on the promise of self-driving vehicles driving profits in the 21st Century are sure to be disappointed.