A Case Study: Self-Driving Cars

Here is Sebastian Thrun’s TED talk explaining why he is developing self-driving (driverless) cars in order to promote safety.

A recent piece in the New York Times by economist Tyler Cowen (May 29, Regulations Hinder Development of Driverless Cars: Can I See License, Registration And C.P.U.?.) discusses the issues with regulation of autonomous vehicles on public roads. While such technology may not be critical in the future, such automated devices have many potential benefits, of which increased safety is touted as a significant possibility. Google has been making significant progress on this front, and the project leader – AI expert Sebastian Thrun from the Stanford AI Lab – claims that its vehicles are already sufficiently reliable even while “safety has been our first priority in this project.” Assuming such claims are reliable for the moment, Cowen observes that current laws are preventing companies like Google from the full potential of innovation in the field, effectively preventing progress towards the goal of tested safety.

Cowen makes the major point that extensive legislative change is slow, and the necessary attention is not high enough on the agenda of current political discourse. There are several issues that make legally responsible entities reluctant to take on the issue, but these mostly stem from the complicated issue of liability and when the safety of human lives is involved. Driverless cars may be safer, but as Cowen ironically states:

“The lives saved by the cars would not be as visible as the lives lost, and therefore the law might thwart or delay what could be a very beneficial innovation.”

Among other issues this arises from widespread innumeracy – or perhaps distrust of good reason in the face of risks in general. A 2010 op-ed from the New York times regarding risks with the Toyota Highlander carefully argues how the irrational public response to major accidents distorts the real risks at hand. While it may seem cold and utilitarian to focus on “the real risks” and dismiss individual tragedies, the sheer frequency of automobile accidents (in contrast to, say, airplane accidents) could see marked improvement if we focused on improving the relevant risks while being mindful of safety – not paranoid.

Although Google is making a very clear effort to emphasize the safety of the technology in this case, development is hindered by legislation to a large extent. If we would like to see the benefits of safe and advanced robotic technology sooner, it would be better for the public and the government to encourage progress than hinder it bureaucratically. Google is developing driverless cars correctly (for which Thrun makes a strong case), this should be welcomed. Pithily, the sooner we test whether such cars are safe, the sooner we may have cars that are safe – and the same should apply to robots in general.