Ever since General Motors rolled out the latest game plan for self-driving cars earlier this month, I haven't been able to shake a line that CEO Mary Barra, among others, offered up: The carmaker's vision of an autonomous future, she said, is for traffic crashes to be a thing of the past.
— Mary Barra (@mtbarra) November 29, 2017
It's noble, sure. Does anyone oppose this idea? Who out there is for more crashes, more emissions, more congestion? Of course, everyone wants zero crashes. Autonomy proponents regularly trot this out as a main goal: by removing humans from the equation, entirely, we can reduce the glaringly high number of traffic fatalities to virtually zero. In May, the U.S. Department of Transportation said as much: Autonomous vehicles are, the department assessed, "driving us to a Zero Death Future."
But when you start to consider the possibilities of what could be done in the near-future, it becomes hard to accept that the zero crash goal is being pursued in earnest.
If the unequivocal goal is creating safer driving environments for millions of passengers on the road every day, there's a lot more that carmakers could - but are not - doing about it. This doesn't even have to include possible future tech, even. There's readily-available safety technology for every new car that automakers could be implementing right now.
GM's ideas - anchored, ostensibly, around this grand vision of safety relying on distant, trickle-down future tech - never seem interested in this, at least publicly. The same could be said about the majority of tech developers and automakers. Why, though? Why not take this sort of action right now? Forward collision warning and automatic emergency braking, for instance, are two features that've been found to cause a 40 per cent drop in crashes. At the moment, it's being rolled out across model lines, often an optional extra.
GM even has semi-autonomous tech in production right this very moment. New Cadillacs can come with Super Cruise, a feature that holds your car in your lane for you, tying together all of the driver-assistance systems we've seen debut over the past few years. GM talks about Super Cruise like a safety system, but it's only available to those who have plenty of cash on hand - right now it's only on the pricey Cadillac CT6 sedan. How long with Chevrolet buyers have to wait before they're permitted to have as much safety as Cadillac customers?
Is the problem that these features cost too much for the average car buyer? GM could be eating that cost if it was truly selflessly devoted to promoting safety, but it's not. And if GM was still forced to pass its development costs on to the customer, why not come out in support of policies that would lead to raising wages across the board? (I'm using GM as an example here, by the way; the same could be said of a lot of automakers pioneering this tech on luxury cars.)
In GM's most recent presentation to investors on the benefits of its self-driving development, the company said that "EV is the foundation for autonomous vehicles." Why? If this really is a life-saving technology, why not work to get it in as many cars as possible. Why confine yourself to electrics, a sliver of the car buying market? You'd think GM's Zero Crashes vision would have the company aiming to get this into as many Chevy Cruzes and Equinoxes as possible. It's not like it's impossible to make a self-driving car with a gas engine. The first car to win DARPA's autonomous vehicle challenge was a diesel Touareg. Google's first prototypes were hybrid. Waymo's current Chrysler Pacificas are hybrids. GM's actions don't match its lofty rhetoric.
There areother ways to bring down traffic deaths, as my colleague Raphael Orlove pointed out late last month.
There are ways that we have access to right now, not dangling perpetually ahead of us - things like making our roads safer, designing cities better, building safer conventional cars and creating smarter crash tests. (Even better driver education would go a long way, in theory, though in practice that's tough to implement too.)
Does any of this fit in the autonomous vision of the future with Zero Crashes?
The Zero Crashes ideal is pitching a drastic change from the status quo, and it requires drastic steps to be taken. This goes far beyond planning new automated models and slowly rolling out tech from the top of the market down. It's not unprecedented in the auto industry for companies to massively overhaul their business.
Ford was presented with such a case in the 1920s, when Model T sales started to slide. Realising it needed to change to prepare for the release of the newer, better Model A, Ford laid off thousands of employees and shut down operations, as Automotive News recounted:
... the wait was painful for those associated with Ford Motor Co. Ford workers were laid off as factories were shut down, suppliers struggled to survive, dealers squeaked by selling used cars, and competitors gained an edge as Ford's sales declined further.
The changeover from the Model T to the Model A was the largest and most costly undertaking in industrial history to that point. Historians estimate the cost between $US100 ($131) million and $US250 ($327) million.
Much of the cost was associated with transferring production from Highland Park, where Ford had built the Model T, to the new Rouge plant, a move that took six months.
That's impossibly drastic, but Ford recognised it needed to change to survive. What General Motors and the Department of Transportation are proposing with their Zero Crashes goal is on that level: a world where car crashes no longer happen. How else can you get there without pursuing an impossibly drastic route?
But in the current environment, both automakers and the regulators who oversee them could be too stubborn enough to consider that sort of proposal. Take BMW.
Back in late 2012, the automaker asked the National Highway Traffic Safety Administration if it could be exempt from an unbelted occupant crash standard if the it instead put seat belt interlocks into all of its cars. In other words, the BMW won't drive unless seat belts are in use.
As BMW put it, they could save thousands of gallons of fuel and pounds of safety equipment necessary to meet the unbelted crash test, so long as they guarantee its occupants were wearing their seatbelt in the first place.
Here was NHTSA's response, from Juan Barnett, who reported from the meeting:
"Removing the protection offered to unbelted occupants would be unprecedented for NHTSA considering unbelted crash test requirements date back to the 1970s"
NHTSA doesn't even want to consider the possibility of "unprecedented" change. Of course, consumer advocates may've viewed the proposal as a step backward for safety - but BMW's argument was that it could actually increase seatbelt usage in doing so. So what happened? Both sides were at a stalemate, and both came out looking bad: NHTSA doesn't budge; BMW, despite saying that seat belt interlocks were ready to go and would make their occupants safer, still doesn't put them in its cars, not without that crash test exemption from NHTSA.
With autonomy, this presents a terribly strange moral conundrum. It's disingenuous to say, "this technology is better than humans and will save lives," and then not do everything in your power to ensure that technology is universally adopted, as soon as possible.
We're not talking about the first wave of expensive cars that, at first, get Apple CarPlay and then, one day, that technology becomes standardised. It's about life or death. Why put that off? Because the markets said so?
According to every autonomy proponent, robots will - one day - overwhelmingly be seen as safer drivers than humans. Why, then, would you not support whatever policy's necessary to ensure that every driver has access to such a vehicle?
There's a solution, however unprecedented it may be, that can achieve GM's vision. When you consider all of the barriers to reach near-zero car fatalities, it's the logical, inevitable conclusion: ban driving. It's not something any car enthusiast wants to hear, and certainly not something I support. But you can see where the thinkers come to this conclusion.
It's jarring to consider, but with one of the largest automaker's in the world pushing such a vision, it has to be contemplated. If GM, the department of transportation, and any developer of self-driving cars, actually wants to reduce the number of traffic fatalities to zero, a ban will, eventually, have to be on the table.
How to go about implementing a ban seems difficult at first, but the basic notion could be this: set a date - say, by 2050 - and do whatever's necessary to ensure everyone has access to an autonomous car. (California lawmakers are already considering banning gasoline cars by 2040, a drastic move in its own right.)
GM can say it wants zero crashes, but that's not an argument. There's no one standing in opposition to that goal. If GM wants that, it can lead the way and produce only fully-autonomous cars and support a ban on driving. If you truly believe that autonomy is the way to achieve zero crashes, how else could you get there?
If GM sticks to its current plan of only trickling out self-driving tech across limited production and high-end cars, a driving ban would cut out low-income car buyers and owners, as as Jalopnik's own Patrick George recently pointed out. That's GM's customer base. Unless GM absolutely commits to getting autonomy into the hands of every driver in America, promoting a driving ban would mean shooting itself in the foot. The only leading face who even suggests a ban is Tesla's Elon Musk. Tesla thinks it can have a fully-autonomous car on the road by 2019, but without the push of someone like GM, there's no question that human-driven cars are going to be in production indefinitely.
If Barra and her colleagues aren't willing to consider a ban, a vision of the future in which there's zero crashes and zero congestion is nothing more than puffed up drivel to sell investors on. Put more simply, you won't eliminate the majority of car crashes unless you ban driving. But we as a society have to decide if this is what we really want.
That's my main problem with the pro-autonomy tagline of a Zero Death Future. GM, Ford, and every automaker developing self-driving cars aren't serious about it. They're serious about being successful car companies. With public opinion still (for the most part) against automated cars, with income inequality continuing to expand, there's no question regular, human-driven cars will be in production for the foreseeable future.
That plays seamlessly into the ambiguous timelines offered by autonomy proponents. When will we arrive at the Zero Death Future? How many years are we talking? A quarter-century? More than 100 years? No one ever says.
That's the problem. Car companies haven't been interested in presenting a realistic timeline, because in order to actually achieve that kind of drastic change being pitched, it would require them to wholly and totally reconsider the way they do business - which is to say they'd have to ditch making human-driven cars. That goes against the grain of their very business model, and car companies exist to run successful businesses. They're not charities. They're not utilities. They exist to make money.
Baked into GM's presentation was an extensive passage by the automaker's president, Dan Ammann, about what it sees as the potentially huge business opportunity for self-driving cars.
"We believe this is the biggest business opportunity since the creation of the internet," Ammann said.
That business opportunity begins with deploying fully-autonomous cars in 2019 for car-sharing purposes. It jibes with most of the industry's vision in the near-future - early on, proponents say, self-driving cars will be intended for limited spaces for Uber-like services. Google's self-driving car unit, Waymo, has a driverless pilot ongoing in Phoenix; it's likely to stay contained for some time. Ford doesn't think it can get an autonomous car on the road until 2021; when that happens, it will be for geo-fenced ride-hailing and ride-sharing.
More and more, as the regulatory landscape and reality itself comes sharper into focus, I'm inclined to believe that self-driving cars are going to be limited to that sort of capacity, indefinitely. And, really, it is a great business opportunity. A fully-autonomous taxi service that doesn't require a driver at the wheel presents a scenario where the owner of the business reaps the entirety of a passenger's fare, and only has to focus on upkeep of the vehicle. The vehicle itself is constantly ferrying people around 24/7; robots don't need to sleep.
It's no surprise that GM's so gung-ho about the idea of self-driving cars as a lucrative business endeavour. Hell, Uber thinks driverless cars are absolutely essential for its business in the future.
And so I'm left to wonder what's the point of all of this. If it's to launch automated taxi services, then cool - say that. If you're willing to take it a step further (as GM is implying with its Zero Crashes rhetoric), and suggest that automation can rid the world of traffic crashes, then actually lay out how you think we can get there.
Otherwise, it's plainly obvious what the end-game is. Ammann said as much: it's the greatest business opportunity since the internet. It's about making money - for now that means ferrying select passengers around by way of automated ride-hailing services. Not safety.