What's the acceptable death rate for autonomous cars?

Before a surgeon cracks open your chest, ask him his acceptable death rate for patients.

When you climb aboard an airplane, ask the pilot her acceptable death rate for passengers.

You'd think — and hope — their answers are "zero," reflecting the overriding philosophy, paramount concern and commitment through which they make decisions that affect your health and safety.

In contrast, the race to bring autonomous cars to market appears to have abandoned this notion of public safety. The overriding industry message is that, once the bugs are worked out, so many more people will be saved by the promise of autonomous cars that killing a few here and there is acceptable in the name of progress.

It doesn't have to be this way.

Many participants in the race to develop autonomous cars have little to no long-term experience or knowledge of the complexities of driving, traffic safety issues or how safety progress has evolved in the transportation sector. There is no shared safety principle guiding the design and testing of autonomous cars. Tesla chastised the media for focusing on the death of a customer whose vehicle's Autopilot software failed, noting that people die in car crashes every day. Yes, people do, but that shouldn't be an excuse to kill more.

History provides lessons and opportunities for advancing autonomous vehicle development in a safer, more coordinated way. The American automotive industry started with little or no safety engineering built into vehicles and vigorously fought the introduction of government-mandated safety standards, arguing it would increase costs and harm sales. Safety efforts, they insisted, should be focused on the driver, or the "nut behind the wheel."

Safety engineering became standard features only after federal legislation in the early 1960s created NHTSA, which issued mandatory vehicle safety regulations. Many regulations were redrafted in the 1990s to be performance-based and encourage greater innovation and use of emerging technology. Vehicles get safer each year, and manufacturers increasingly compete on safety.

Automakers and suppliers have a wealth of experience and knowledge based on billions of miles driven, hundreds of millions of cars built and reams of data compiled from vehicle crash tests, consumer and dealer feedback, crash investigations and safety recalls. The basic premise is to test and ensure the safety of vehicle designs before you place them into the marketplace. That's why automotive test labs and proving grounds exist.

The race to bring autonomous cars to market has turned this process upside down. State governments, tech companies and car manufacturers are rushing forward to be first to market, leaving safety last. Autonomous vehicles will have to be tested in our streets to truly learn and improve, but that step should occur after passing sequential safety milestones, not before. Simply using our communities as the testing grounds, and our families and children as the crash-test dummies, is not only wrong, it's nuts.

As an emergency physician, I know that not every injury can be prevented, nor can every crash victim be saved. But the overriding principle in medicine, the oath that every doctor takes, is that first, we shall do no harm to our patients.

Autonomous cars can provide tremendous benefits in terms of safety, environment, health and mobility. Death by car is still a leading killer in the U.S. and worldwide. In order to achieve the benefits and build public trust, there must be a strong unifying vision, a shared commitment to safety and a coordinated effort by governments and the industry to make autonomous vehicles safe for all. This is not only possible; it's being done in communities around the world.

One example is Vision Zero, a philosophical concept that started in Sweden in 1990 and is based on two radical ideas that now are well-accepted. First, every injury and every death on the roadway is unacceptable. Second, the designers and developers of roadways and vehicles would take responsibility for the safety of their systems. This shared mindset has brought diverse groups together around a shared vision and enabled them to achieve impressive results in reducing traffic deaths.

In the U.S., different government agencies, technology companies and automotive manufacturers have varied approaches to Vision Zero, but none have decided to truly "own the problem." If you disconnect the driver from control of the vehicle, doesn't somebody have to step up and take responsibility? Physicians and nurses do it every day.

Owning the problem, taking responsibility and being accountable are essential to building the public trust needed for success.

Another example is the Health Effects Institute, a public-private partnership that seeks solutions to potential and recognized health risks in the transportation sector. It represents the type of independent leadership required to bring all sides to the table, secure the public trust and ensure that the U.S. remains a world leader in development of autonomous transportation.

Convincing major players in the intensely competitive automotive and technology sectors to collaborate on a shared goal of safety as the top priority won't be easy. Asking them to take responsibility may be even harder. But the opportunity to accelerate the growth, acceptance and success of autonomous cars and American innovation requires it.

Skeptics say, "You can't really prevent all motor vehicle deaths and injuries, can you?" The answer, of course, is no, but you can try and commit to adopt the philosophy that one will, first, do no harm.