Home / Blog / Autonomous Vehicles: Myths according to Philip Koopman

Autonomous Vehicles: Myths according to Philip Koopman

As regular Critical Link blog readers know, I’m both a car guy and a technology guy. So I am naturally drawn to articles at the intersection or car and tech. Thus, my interest in autonomous vehicles (AVs). EE Times recently had a provocative article by Carnegie Mellon associate professor Philip Koopman, whose specialty is AV safety. While Koopman definitely sees the potential benefits of AVs, he argues that AV technology is still “immature”, and that AV industry campaigning to keep regulatory oversight at a minimum is wrong-headed. This campaign, Koopman says, is based on what he calls “The Dirty Dozen” myths about AVs – a list of myths he lays out and debunks.

Using Koopman’s myth statements in full, I’ll briefly summarize his points, but urge you to read the full piece. Even if you don’t agree entirely with his premise that we need more AV regulation, not less, overall his points hold a great deal of validity.

Myth #1: 94 percent of crashes are due to human driver error, so AVs will be safer. The gist Koopman’s argument is that this misrepresents the original findings of a study that “94 percent of the time a human driver might have helped avoid a bad outcome.” Which is not exactly the same as accidents caused by the driver. Drivers do make mistakes; so will AV. It’s just that their mistakes will be different. AVs won’t be texting while driving, but there’s plenty that can go wrong.

Myth #2: You can have either innovation or regulation – not both. In fact, Koopman argues, you can have both. And that regulations can be written that explicitly allow for innovation. Even if the industry just sticks to its own generally agreed upon standards for safety, training, and testing, there’ll be plenty of room to innovate.

Myth #3: There are already sufficient regulations in place. Other than in NYC – where there are an awful lot of cars on the streets and pedestrians in the crosswalks – Koopman says there are precious few regulations. When it comes to safety, regulation equates to little more than taking the manufacturer’s word for it.” Gulp!

Myth #4: We don’t need proactive AV regulation because of existing regulations and pressure from liability exposure. Not true, Koopman says. The regulations covered under today’s Federal Motor Vehicle Safety Standards are principally focused on basic safety functions – think headlights and seat belts – and do not address computer-based system safety. Which is pretty much what AVs are about. With respect to liability exposure, Koopman’s wary that companies with deep pockets for development may not mind making a few payouts (however large) to the families of those killed by an AV.

Myth #5: Existing safety standards aren’t appropriate because (pick one or more):

    • They are not a perfectfit;
    • No single standard applies to the whole vehicle;
    • They would reduce safety because they prevent the developer from doing more;
    • They would force the AV to be less safe;
    • They were not written specificallyfor AVs.

Hogwash, according to Koopman. There are existing safety standards that are flexible, and which cover AVs. Nothing in these standards would keep a developer from doing more than required and none of them would render AVs less safe (an argument Koopman finds “laughable”).

Myth #6: Local and state regulations need to be stopped to avoid a “patchwork” of regulations that inhibits innovation. These “patchwork” regulations exist in large part because AV companies pressure the state’s they’re working in to go easy on regulations, threatening to pack up and take their jobs to a more accommodating state. “Moving to regulation based on industry standards would help the situation. A federal regulation that prevents states from acting but does not itself ensure safety would make things worse.”

Myth #7: We conform to the “spirit” of some standard. Spirit of the law vs. letter of the law arguments may seem like hairsplitting, but this makes Koopman’s point pretty clearly:

Consider whether you would ride in an autonomous airplane in which the manufacturer said: “We conform to the spirit of the aviation safety standards, but we’re very smart and our airplane is very special, so we took liberties. Trust us – everything will be fine.”

Myth #8: Government regulators aren’t smart enough about the technology to regulate it. There is no denying that the technology underpinning AVs is complex. It’s not really a matter of smart vs. not so smart. It comes down to technology that is rapidly changing and requires tremendous expertise to truly understand. But Koopman points out the plan being proposed by the U.S. Department of Transportation invokes the standards that the AV industry has already agreed to. Which sounds smart enough to me.

Myth #9: Disclosing testing data gives away the secret sauce for autonomy. Nope, Koopman says. Reporting testing data won’t disclose a company’s intellectual property.

Myth #10: Delaying deployment of AVs is tantamount to killing people. So far, the safety benefits of AVs are all in the long-promised future. And so far, there’s nothing to prove that AVs will be safer than cars driven by humans, especially given that those human-driven vehicles are getting safer thanks to the growing use of active safety features (e.g., automated emergency braking).

Myth #11: We haven’t killed anyone, so that must mean we are safe. Hmmmmm. I just googled and it seems that there have been at least a handful of deaths associated with AVs. Not to mention that ‘so far, so good’ doesn’t equate to perfect safety.

Myth #12: Other states/cities let us test without any restrictions, so you should too. Road testing is important, and no one’s going to trust AVs until and unless they’re fully road-tested. But just because one jurisdiction allows for it, that doesn’t make it safe to do so. Here, Koopman gets back to his central theme that the AV industry really just has to adhere to its own agreed to safety standards and not put innocent drivers and pedestrians, caught up unwittingly in someone else’s experiment, at risk.

As disappointing as it’s been that AVs aren’t further along by this point, I’d hate to see things rushed to the point that there are enough public failures that the risk seems to start outweighing the rewards of autonomous vehicles. This could result in a backlash – and some over-reacting, over-regulation – that could stymie advances for years to come. I’d hate to see that happen. Sounds like some regulation is in order.