By Tarun Wadhwa
Published: Mar 22, 2018 11:12 a.m. ET
There are many things we still don’t know about the self-driving Uber car that killed a pedestrian in Tempe, Ariz., on Sunday night.
But we do know this: the technology and car companies need to be more honest with the public about the severity of challenges they face in developing autonomously driving vehicles.
There will surely be many more crashes, smaller incidents and problems along the way. We haven’t even begun to scratch the surface of the cybersecurity challenges of having vehicles safely and constantly communicate with each other and everything around them. We don’t know how to deal with the national security implications if these machines are misused by terrorists. And we have yet to have an open conversation about the loss of the privacy of our movements, or what it will mean to rely on companies for life-or-death decision-making as we move around.
We must start with a hard look at how we’re deploying this technology and whether it is being done the right way.
Today, 21 states that have laws regulating partially-autonomous vehicles, with Arizona taking the lead by setting lax standards and allowing regulatory guidance to come after the fact. Contrast that with California, where Uber once dangerously deployed self-driving cars into San Francisco, against the government’s wishes, that lacked the ability to properly detect bicyclists, among other things. A car ran a red light in a busy intersection. Uber blamed the driver but evidence later showed the car had malfunctioned. After several days, the state forced Uber to stop its experiments with self-driving cars on city streets.
We have seen with Amazon’s search for a new headquarters just how desperate politicians are to attract technology talent and how much they are willing to give away. We can’t have a similar process for self-driving cars. The stakes are far too high, and we are incentivizing the wrong things. We need to have companies build and design their systems at the highest possible standards, because driving is full of unknowns. By allowing them to instead meet a minimum threshold, we are undermining public trust and slowing down the technology’s development.
Governments need to develop ethical frameworks for companies to follow. This is not something that can be left to the private sector to develop. We need better, more open and public tracking of accidents. Today, there are many low-level incidents — fender-bender and near-misses — that get little attention.
As we adjust our laws to reflect the existence of self-driving cars, we need to tackle the tough philosophical questions over who is responsible when there is a crash and, even more, when someone is injured or killed. Is it the user, the software maker, the automobile maker, some other part of the supply chain or a combination of them?
Without resolving these questions, we cannot realize the positive benefits that autonomous vehicles can offer.
There are dozens of companies — from General Motors GM, -2.90% and Alphabet GOOG, -3.18% GOOGL, -3.13% unit Waymo to less prominent firms — working independently to develop driverless cars. At stake isn’t just how people will get from point A to point B, but how future societies will work, what mobility means as a quality-of-life issue, and what choices that we as individuals will be able to make about how we live.
The automobile changed many things about our society — how we travel, how we eat, how we shop. Self-driving cars will do all of that, plus a lot more. We will be able to reclaim billions of parking spots, countless hours lost in commuting, change our relationship to the land around us. That includes bringing traffic injuries and deaths down significantly. And it isn’t just cars: anything that moves will eventually be touched by the same underlying technologies and one day be self-driving and self-thinking.
We are still years away from full autonomy, but we must continue trying to get there. The innovations we create along the way can save countless lives. In 2016, 37,461 people died in motor vehicle crashes in the U.S., according to U.S. Department of Transportation data. Let’s keep in mind why we are starting down this path: because driving today is too dangerous, and the potential public safety gains from of autonomous vehicles are massive.