Will self-driving cars ever be safe? Unfortunately, the first pedestrian fatality involving autonomous vehicles just occurred in Tempe, Arizona. And it has cast a long shadow across this burgeoning industry once viewed as an inevitable part of our future.
Now, everyone involved is left to pick up the pieces and answer the one question on the general public’s mind — is this something still worth pursuing?
Comparing Apples to Oranges
The above question isn’t just for AI developers, autonomous car engineers, or the investors funding these initiatives; it’s a question for everyone. Because self-driving cars are poised to play such an integral role in our future society, it’s vital that we all consider the repercussions as well as the benefits.
To start, it’s crucial that we view all facts fairly without influence or forced perspective. In response to this recent tragedy, many social media users were quick to point out that 37,000 people lost their lives due to human-driven vehicles in 2016. Putting this up against the one recent pedestrian fatality makes it seem obvious that AI-piloted cars are doing pretty well. But this is actually an overly-simplified way of skewing the data.
To be objective, we have to compare both categories in the same units of measurement. For instance, human-piloted vehicle fatalities are measured in “vehicle miles traveled.” For 2016, this equates to roughly 1.18 fatalities every 100 million miles driven by Americans. Americans drove approximately 3.2 trillion miles in 2016. This still surpasses 30,000 fatalities.
Now, let’s do autonomous vehicles. While they looked great before at a superficial level, the truth is, they’ve driven far less than 100 million miles before this first accident. Uber had just reached 2 million miles in its self-driving program. Waymo recently logged 4 million miles for its initiative. And we won’t count Tesla’s autopilot because it’s not completely autonomous. So adding up all of these miles actually makes autonomous driving look very bad compared to human-piloted vehicles.
It’s Never Been About the Numbers, It’s About the Potential
But before you get your pitchfork, it’s important to remember why so many organizations have embarked on the endeavor to make autonomous vehicles. The potential — what could be if we get it right — is what caused so many people to pour so much time and money into this effort.
Self-driving cars could change transportation from the bottom up. Traffic jams in cities like Los Angeles would be non-existent. Hours you spend driving every week could be put to use in a variety of other productive ways. The elderly who are vastly unequipped to still operate a vehicle could get to where they need to go in a much safer way. Drunk driving accidents could plummet.
It isn’t fair to pass judgment on something in the middle of its creation. That would be like taking a cake out of the oven halfway through its baking time, deciding it tastes bad, and trashing it.
A great example of this actually happens to be the release of the regular automobile as we know it. In 1921, every 100 million miles traveled brought 24 deaths. It may have taken some time, but that figure dropped to only 1 because of persistence and constant refinement. That is exactly what self-driving cars need.
Assessing the Situation for What It Is
Clearly, how you view self-driving cars to human-piloted cars dictates your judgment of the two. But to more accurately decipher how dangerous self-driving cars really are, we’ll need them to drive billions of more miles. They don’t have enough miles on the road to know how dangerous they really are, just that they can be dangerous.
This isn’t to say that they will most certainly end up being safer than human drivers. After more miles are logged, we may find ourselves with other obstacles facing self-driving cars that we never envisioned. But only time on the road will elucidate these problems, and the sooner we do, the sooner we can start working on solutions.
Being Right Is Better Than Being First
Uber has pulled the plug on its public autonomous car venture in light of this tragedy. If the company hadn’t, the immense public pressure probably would have made it in a few weeks from now. This unfortunate accident may have done lots of damage to the reputation of self-driving cars, but it wouldn’t have been possible if autonomous vehicle enthusiasts and companies weren’t acting like the technology was already here to use.
Self-driving cars are still very much in a state of experimentation and refinement. It’s likely that many tech companies are putting their bottom lines before the public interest and see this whole thing as a race to get to the money first. Let’s recall the moral of The Tortoise and the Hare: it’s better that we take our time with this and get it right rather than rushing to the finish line. And if this means taking self-driving cars off of public roads and revamping them, then so be it.
There are some races that nobody wants to be the winner of, and unfortunately, Uber just became first in this one.Tags: AI, AI App Developer, AI apps, artificial intelligence, artificial intelligence app developer, artificial intelligence app development, autonomous cars, autonomous vehicles, Los Angeles, los angeles app developers, Los Angeles app development, self-driving auto, self-driving car technology, self-driving trucks, self-driving vehicles, uber, uber app, uber controversy