Northbound on the MoPac Expressway with the sun low in the sky and jazz playing quietly on the satellite radio, Michael activates the Autopilot feature in his Tesla Model 3. He keeps his hands on the steering wheel and feet on the pedals, but he allows his entire body to relax. The car slows almost imperceptibly to adjust to the flow of traffic, and the screen affixed to the center of the dash notes the positions of nearby vehicles in relation to him.
A tone alerts him to a new fare in the vicinity. Having already eaten dinner and not planning to turn in anytime soon, Michael has been aimlessly cruising the city and waiting for notifications from one of the three ridesharing services—Uber, Lyft and RideAustin—that he drives for. Passengers are usually excited to ride in a Tesla, and most of them are fascinated by the computer display showing the traffic around them. It’s not uncommon for someone to ask him if he’d mind turning on the Autopilot. He likes to think that the cool car helps him to net better ratings and tips, although he wishes he could work out a better way of monetizing it.
The funny thing to him is that the vaunted Autopilot feature really isn’t a game-changer for him as a driver. It provides him with a pleasant opportunity to lower his guard and let the stress from the day melt off of him, but he doesn’t miss it terribly when he’s driving a rental car or one of his day job’s corporate vans. Its main function, as far as he’s concerned, is to improve his image.
From an actual driving standpoint, it’s not a whole lot different from cruise control.
As an ambivalent user of self-driving auto technology, Michael likely doesn’t grasp his own importance as a consumer. Ten or 15 years from now, a newer model of Tesla may be ferrying customers around town by itself, and he’ll be left to find a new side hustle. But the tricky thing is that the car manufacturers and software engineers are going to have a hard time making drivers like Michael redundant without help from the Michaels of the world themselves.
As with many anticipated technological breakthroughs, you could be forgiven for thinking that self-driving cars are either here already or right around the corner. According to one survey, 71 percent of drivers believe that they could go out right now and buy one (and 11 percent think it might be fine to take a nap using the tech currently available). In fact, there is no car on the market capable of handling a typical morning commute all by itself.
One of the challenges involved in discussing “self-driving” or “driverless” automobiles is that there is plenty of middle ground between total human control and total computer control. Technical write-ups usually rely on the classification system devised by the Society of Automotive Engineers, which lays out six levels of autonomy. Level 0 leaves everything up to the driver, while Level 1 includes relatively common features like adaptive cruise control and parking assistance. Tesla’s Autopilot and Nissan’s ProPilot are considered Level 2, as the car steers but the driver keeps hands on the wheel. Audi has introduced Level 3 in several models, which is intended to allow the driver to engage in other activities while still being ready to take over if necessary; this currently functions with a maximum speed and is primarily designed for use during traffic jams.
Levels 4 and 5 are currently not available in any mass-produced automobile. Both of these refer to total computer control of the vehicle, with 4 limited to particular routes or weather conditions and 5 entirely autonomous. Manufacturers aren’t in agreement as to whether 5 is ever going to become a reality, although Tesla CEO Elon Musk has boasted that his brand will achieve 5 by 2020. Waymo CEO John Krafcik, on the other hand, has stated that “autonomy will always have constraints.”
Moving goods around a small, well-mapped area will probably be the first major function of truly driverless vehicles, with Kroger already running a pilot program for automated grocery delivery in Scottsdale, Arizona. Ferrying live passengers is a very different matter, which will (and should) be approached with caution by everyone involved. While much of the reporting on autonomous vehicles warns us to get ready, actually getting there is going to involve clearing some considerable hurdles for the auto industry.
While the technology for at least Level 4 autonomy is feasible, there is no guarantee that widespread adoption is inevitable. Many transportation advances, from the monorail to personal rapid transit, have appeared full of revolutionary potential but failed to become commonplace. Driverless cars offer the prospect of greater safety because they don’t get distracted or angry like human drivers, but consumers commonly overestimate their own skills and may have a hard time ceding control to a computer. If not properly regulated, self-driving cars could increase air pollution, and their collision-avoidance systems present a range of ethical difficulties.
Add to this the widely reported deaths of a driver in Florida in 2016 and a pedestrian in Arizona in 2018, and the forecast for self-driving vehicles is not as sunny as the media hype would have you believe. “We overestimated the arrival of autonomous vehicles,” Ford CEO Jim Hackett mused recently. Back in 2016, a report by UBS estimated that widespread adoption of driverless vehicles was “decades away,” due to the amount of testing needed to demonstrate that the technology is safe.
Getting to Level 5, or even Level 4, will still take some technological development; more importantly, it will require a dramatic transformation of society’s attitudes toward driverless cars. In the short term, the audience whose buy-in is most needed is not what you might expect.
Reaction to the prospect of driverless vehicles tends to be split into two highly vocal camps: those who can never see themselves willingly relinquishing control of the road to a computer, and those who can’t wait to nap or play video games on their way to work. (For a more nuanced glimpse at the range of attitudes, see the survey in this week’s newsletter.) Once Level 4 or 5 technology is readily available, auto manufacturers can concentrate on cultivating brand loyalty among the enthusiasts and winning over the die-hard drivers. But neither one of those camps is going to provide the crucial help that the industry needs to reach this tipping point.
The people who are most excited about autonomous vehicles are often interested in the labor-saving aspect of the technology, but it will be some time yet before drivers are able to safely take their attention off the road. Driver inattention was a major problem encountered by Waymo in a Level 3 test project using employees of its parent company Google. “After looking at the data from onboard cameras, they stopped the program,” said Waymo advisor and former General Motors R&D vice president Larry Burns. “People were falling asleep, eating, reading—they were doing things that made it impossible to re-engage the driver.”
As a result of these findings, Waymo essentially turned its attention to Level 4 autonomous vehicles and bypassed Level 3. But for a manufacturer like Tesla, whose business model depends on getting the most cars out on the road, skipping intermediate steps in order to focus on long-term results is not a viable option. So how do you sell drivers on expensive new technology that won’t actually do the bulk of what they want it to do?
The answer involves reaching out to drivers like Michael—consumers who have the purchasing power to afford the latest automotive bells and whistles, but who approach self-driving technology with a degree of ambivalence. To put it bluntly, these drivers will be willing to engage with semi-autonomous “driver assistance” superficially, as a showroom feature along the lines of a heated steering wheel or built-in vacuum cleaner. The Michaels of the world see Autopilot and its analogs as fun gadgets and status symbols, not as something to help them reclaim their commute time.
This is an odd and uncomfortable moment for automotive brands working on autonomous vehicles: They can’t win over reluctant drivers until the technology becomes reliable and commonplace, but they can’t perfect the technology without getting it onto the market (and hence generating data and making mistakes) in its current, not particularly useful form. Getting to Level 4 or 5 will require a completely different marketing strategy than what will come afterward.
So where does the industry locate its Michaels? Early adopters are likely to be found in the corporate and government worlds, with taxi companies, corporate fleets, airport hotels and city bus services likely to be interested in the passenger-facing value of driver assistance features. Customers and employees can take comfort in the fact that the companies have spared no expense to ensure their safety and avoid driver error. The technology will function less as a specific feature and more as part of a carefully constructed brand image.
As far as individual consumers, car buyers who supplement their regular income by driving for ridesharing services are a safe bet. In order to maximize the marketing potential of self-driving tech, auto manufacturers should look into partnering with ridesharing companies. Uber and Volvo are already working together on autonomous vehicles for the future, but they could also collaborate on a pilot project in which selected drivers are loaned cars with driver assistance features that are announced to passengers within the app. Michaels don’t have to be Uber drivers, either—anyone with clients to impress is likely to be interested in acquiring the most cutting-edge automotive features. Salespeople, realtors, tour guides and more can find a way for self-driving technology to enhance their professional aura.
The design ramifications are that while user-friendly interfaces for features like Autopilot are key, it can’t hurt to play up the sexy, high-tech looks and sounds. It might feel tacky to treat complex, potentially life-saving systems as glossy luxury features, but doing so is the best way to ensure their long-term viability.
It might be asked whether Michael should take part in supporting a technology that is likely to make his own driving skills obsolete and detract from his income. Individual consumers will have to make this choice for themselves, but there is no reason why ambivalent drivers like Michael can’t become enthusiastic adopters as self-driving cars evolve to become truly autonomous. What makes them the ideal consumers at this stage is not their ambivalence itself, but their lack of expectations about what autonomous cars can do for them. By using the technology responsibly and popularizing it, they will provide the basis for its refinement—and possibly win over a few of the anti-autonomous drivers along the way.
Amidst the eager anticipation of the techno utopians and the stubborn refusal of the die-hard drivers, the Michaels of the world aren’t making their voices heard. And car manufacturers have to worry so much about the long-range trajectories of their inventions that it can be hard for them to confront the problems that are right in front of them. Recognizing our current moment as a critical transitional period for the auto industry is what will allow them to target the right consumers.
Getting the next few years right will determine which brand, if any, emerges to become leader of the self-driving pack. And if none of them play their cards right, it’s entirely within the realm of possibility that the much-hyped driverless revolution simply won’t happen.