The Right Path to Autonomy—and the Wrong One

Self-driving cars? Not until the friction between morality and culture are resolved.


What is the best path to autonomy? I was recently asked to join the debate over at 2025AD, an excellent site focused on the year many claim will be the watershed for the arrival of self-driving cars. I'm bullish on the morality of autonomy, but bearish on current methods of getting there. My opponents? Most of the pro-autonomy community.

As a driving enthusiast, people often ask me why I’m so enthusiastic about self-driving cars. The answer seems obvious to me, but it’s not because I’m aligning myself with the pro-autonomy camp. The benefits autonomy will bring would appear to resolve a universe of problems that are seemingly intractable without it, but I think both sides are absolutely wrong in their either/or vision of when and how autonomy will change society as we know it.

In that binary vision, each side ignores the fundamental truth underlying the other, and the mutually beneficial co-existence that will only be possible if the fundamentalists on both sides get out of the way.

I have no doubt that a Zero-day—the day when the last human will get into a car and choose to drive on a public road—is coming, but that day is so far off as to be irrelevant today. How far off? At least 50 years, as many as 100.

I also don’t believe in a global, regional or even a national tipping point toward autonomy. Any profound shifts must take place at the intersection of politics and culture, which vary widely even within nations. The United States—one of the most important markets and the de facto R&D lab of the world—is a patchwork of nation-states, and if all politics is local, any technology that relies on politics will be as well.

Between now and that Zero-day—which I first wrote about in the 2015 Jalopnik article, “How Science Fiction Failed Us”—we who hope to see the broader benefits of autonomy face a long slog. For those who don’t see any benefit to direct human control of vehicles, disappointment awaits. No timeline will be fast enough. Why? Because even if a Level 5 self-driving car was made available tomorrow, human nature remains a force of nature no less powerful than oceanic currents. For those skeptical of autonomy, every autonomous mile driven is a dangerous assault on reason and undefined notions of freedom. If society is to reap the true benefits of autonomy, both sides need to understand what the other is trying to achieve.

It is impossible to argue the merits: traffic, pollution and road death rates will improve dramatically—hence the unquestionable moral primacy of autonomy. That is why the human driving camp has only two arguments, one simple, the other they are barely able to articulate: 1) autonomous technology isn’t ready and/or never will be, and 2) “loss of control” or agency.

The technological hurdles will be crossed; that is inevitable. That leaves the issues of control, agency and notions of freedom which will take a lot longer to address. Note that I said address, not solve. Technology can solve a variety of problems, but human nature isn’t one of them. Human nature isn’t a problem, but a fact. Culture can change, but human nature remains a constant within it.

Because the two sides in the autonomy debate are arguing past each other — morality on one, philosophy on the other — there is no single path to autonomy.

There is a right path, and a wrong one.

The wrong one lies in the premature and inaccurate marketing of technologies that can’t be successfully demonstrated, and the skepticism, disappointment and deaths that will yield. Every time a journalist reviews a semi-autonomous car with Self-Driving, Driverless or Autonomous in the headline, the case for autonomy is harmed.

The wrong one also lies in mandating autonomous zones or lanes. Because driving is a privilege rather than a right, any encroachment—even perceived—will lead to resentment, active opposition and the delay of inevitable outcomes. People need to have choice. That’s why social engineering always fail. People will be driving for a long time.They need to opt-in, not resent the lack of an opt-out.

The wrong one lies in believing autonomy is entirely good or entirely bad, and arguing for/against based on technologies that 99.9% of people even within the automotive and tech sectors have never experienced, and among whom the second-order consequences are barely understood. (Here’s a good take from Benedict Evans.)

Sadly, the loudest supporters of autonomy are the ones doing it the most harm. Their dogmatic bleating is as foolish as those who refuse to give up their steering wheels. The majority of mobility “experts” — most of whom know little more than what they read in half-baked studies or see at trade shows —  are doing a great job of inspiring opposition to the very future they hope to see. The history of “experts” is one of mountains of data, guesswork and bandwagon jumping. From the Rand Corporation’s application of system theory in the Vietnam War to the big consulting companies’ failure to predict Uber or Tesla, experts have been wrong again and again. “Experts” are essentially reactive. Garbage in, garbage out. There can be no war on pollution, traffic and road deaths without recognizing that human nature cannot be quantified, and that autonomy is not weapon or even a solution, but a catalyst.

The right path must acknowledge both sides and deploy autonomy as soon as possible within the framework of people’s real wants and needs rather than some utopian vision.

In the near-term, stop trying to convince people who will never want autonomy that it’s coming, or even good. Limit the marketing of Level 4 — as Voyage is doing — to use cases that sell themselves.

In the mid-term, the right path must acknowledge that autonomy isn’t binary, but a continuum. Between now and ubiquitous Level 4+, myriad semi-autonomous technologies yet to be seen will hit our streets, and they aren’t Level 3 as popularly defined. Sorry, but Level 3 as we know it is not the bridge to 4, and anyone developing, deploying or marketing 3 as such is missing a multi-decade opportunity. Augmented driving, currently touted publicly only by Toyota’s Research Institute (TRI), is the ONLY method of attracting the current and next two generations of drivers to any form of autonomy. (TRI’s Gil Pratt podcast interview here).

Augmented Driving is the opposite of L2/3, and resolves the conflict between culture and autonomy. The uncrashable car under human control appeals to the human need for freedom and agency in a way no iteration of L3 can. (Deeper explainer here.)

The right path to autonomy is the shortest one that deploys autonomy in all its forms — including the hybrid human/machine forms that will augment human skills rather than replace them — with the support of all.

Alex Roy is Editor-at-Large for The Drive, Host of The Autonocast, co-host of /DRIVE on NBC Sports and author of The Driver, has set numerous endurance driving records in Europe & the USA in the internal combustion, EV, 3-wheeler & Semi-Autonomous Classes, including the infamous Cannonball Run record. You can follow him onFacebook, Twitter and Instagram.