Ford CEO Jim Hackett Tuesday joined the growing ranks of vehicle and tech execs willing to say publicly that self-driving cars won’t arrive as soon as some had hoped.
The industry “overestimated the arrival of autonomous vehicles,” Hackett told the Detroit Economic Club. Though Ford is not wavering from its self-imposed due date of 2021 for its first purpose-built driverless car, Hackett acknowledged that the vehicle’s “applications will be narrow, what we call geo-fenced, because the problem is so complex.” Bloomberg earlier reported the comments.
Hackett is the latest high-ranking industry insider to engage in public real talk about the prospects for self-driving cars, which back in 2016 seemed just around the corner. Less than three years ago, Uber showed off a limited self-driving service in Pittsburgh, Google spun out its driverless vehicle program into a company called Waymo, and Nissan was sticking to its 2013 promise (by a very confident Carlos Ghosn) that it would debut a self-driving car by 2020. Also in 2016, General Motors acquired the self-driving tech company Cruise Automation for a reported $580 million. And then-Ford CEO Mark Fields said the company would have thousands of fully autonomous vehicles available to urban car-share and ride-hail fleets by 2021.
Now, Hackett and others are acknowledging that the industry’s marketing—and the media’s headlines—may have gotten ahead of the technology. These days, execs are making fewer and fewer promises about when self-driving will arrive.
In November, Waymo CEO John Krafcik said “autonomy always will have some constraints,” and suggested that a self-driving car that could truly go anywhere might never materialize. Though Waymo’s limited self-driving service in suburban Phoenix is slowly expanding locally, the company still puts operators behind the wheel, for safety. “Driverless” is a stretch. The chief technology head at Nissan’s Silicon Valley development center now publicly pooh-poohs the concept of completely driverless vehicles, insisting last month that “an autonomous system without a human in the loop” is “a useless system.” (Former CEO Ghosn has been arrested again by Japanese authorities.)
After a crash during testing that killed an Arizona woman last year, Uber is back to limited testing in Pittsburgh, with two operators on board. In a Reuters interview this week, Raquel Urtasun, the chief scientist in charge of Uber’s self-driving efforts, called the question of when driverless vehicles will arrive “the billion-dollar question.” “And the first thing I learned is no timelines, right?” she said.
Urtasun echoed a sentiment increasingly popular among self-driving developers: It’s not when self-driving cars will arrive, but where. Right now, the technology is designed to operate in specific contexts—the wide, sunny roads of Arizona, or the chaos of Boston, or the sunlit streets of Miami, or the sprawling suburbs of San Jose. As if she anticipated Hackett’s comments, Urtasun said, “self-driving is going to start in small areas and then we’re going to grow from small areas to everywhere. The challenge is to make that transition as smooth as possible.” The upshot: Unless you’re in one of those small areas, don’t hold your breath for a self-driving car.
By contrast, that other Detroit carmaker, General Motors, maintained as recently as November that it’s “on track” to roll out an autonomous vehicle this year, though it has not said where its first, limited, service, will be. The company declined to comment on Hackett’s statement.
What’s so complicated about full self-driving? For one, there are no federal regulations for the tech, and states have struggled to fill the void with their own testing rules. Second, industry insiders say sensors need to get better—to “see” farther more cheaply—before the tech can be deployed widely. And developers are still hacking away at better algorithms, ones that can handle the uncertainty of new road situations without hurting their cargo.
Ford, meanwhile, says it still plans to deploy an autonomous vehicle is some commercial service by 2021, in more than one US city. Along with Argo AI, in which it has a majority stake, the Detroit carmaker is testing in Michigan, Miami, and Pittsburgh. Hackett has also outlined his plan to create a wider “mobility platform”—that is, an operating system that would help cities manage self-driving cars when they do arrive.
There’s one self-driving outlier that has not let up on the hype, and that’s the ur-disruptor of Silicon Valley: Tesla. Less than two months ago, the company again began selling its “full self-driving feature,” a $5,000 add-on that it had removed from its website in October. CEO Elon Musk said the feature would be “complete” by the end of the year, and fully available to customers by the end of 2020, in theory allowing Tesla owners to switch on Autopilot and watch their vehicles drive themselves through complex city environments—or, depending on their fancy, to fall asleep at the wheel. The CEO has a record of missing his own deadlines, but Tesla also usually moves against the herd.