If you’ve watched a news report or read a newspaper lately, it might seem like we’re nearing a golden age of self-driving cars that will revolutionize transportation as we know it. That line of thinking is great for headlines and driving traffic to websites, but it’s quite far from the reality of autonomous vehicles. This raises a bunch of questions, the biggest of which is “why don’t we have self-driving cars yet?”
As it turns out, making a vehicle that can operate itself, outside, in the real world, without killing its occupants and everything around it, is hard. That difficulty is compounded by layers of legal, financial, and even psychological challenges. All of those things add up to a scenario where true self-driving cars are years, if not decades, away. To get a better feel for why that is, I went to Bryan Reimer, an MIT research scientist and the Associate Director of the New England University Transportation Center.
What follows are some of his thoughts on why we’re unlikely to see self-driving cars anytime soon.
Uber has pushed into autonomous-vehicle testing in the hope that it can one day eliminate drivers from its business model, thereby increasing profits. (Photo: Getty Images)
No surprises here. Businesses exist to make money, and in most cases won’t keep doing things that don’t yield financial results. Highly automated technology, says Reimer, needs to be robust, tested, calibrated, and validated. That’s all before it’s packaged and sold to anyone, and our litigious society makes it hard to release such innovations into the wild.
The costs associated with developing and marketing self-driving anything are massive, so the return on investment has to become very real, very fast for companies to remain interested.
Self-driving shuttles that operate in narrowly defined areas will be the first autonomous vehicles to be deployed. (Photo: Getty Images)
As you read this, no one has developed a highly robust Level 4 autonomous driving system that is ready for commercial deployment at anywhere near the scale required to generate revenues that offset the costs of development. We’re far enough away from an autonomous future that Reimer believes there’s nothing, even on the horizon, that can be efficiently implemented.
We see plenty of upstarts putting out autonomous “pods” or mass-transit systems, all of which operate in a geofenced (controlled) area, but none of them even slightly resemble the self-driving cars that we all thought we’d have in 2020.
Autonomous vehicles could contribute to increased traffic congestion, just as ride-hailing services with drivers have. (Photo: Getty Images)
Our country’s leadership can’t even seem to agree on what time of day it is, much less complex and wide-reaching policy decisions, so it’s not surprising that autonomous vehicles remain a legislative mystery. Reimer says that highly deployed automated vehicles can actually increase traffic problems as they circulate through urban areas looking for their next “fare,” but congestion is just one issue that legislators will have to face.
Assuming we reach a point where things like school buses are automated, who will monitor the pickup and drop-off of our children, to regulate the chain of custody? Who will be responsible in the event of an accident? Further, an autonomous vehicle operating in New York state will need to face the same set of laws when it crosses into Connecticut, which has different traffic laws in many cases.
When we give up control of a vehicle to an autonomous service provider, who is ultimately responsible? (Photo: Waymo)
One of the biggest challenges for many people is giving up control, even more so when giving up that control means handing over the keys. A quick look at the drunk driving statistics in this country will tell you that people are hell-bent on getting home behind the wheel of their own vehicle at almost any cost. Why, then, would any one of those people get into a vehicle that not only takes away control, but does so without another human in charge?
As unlikely as it sounds, Reimer says that humans are actually pretty good when it comes to driving. It turns out that we’re well adapted to the multi-faceted challenges of driving and tend to be nearly as good as computers – even better in some situations.
We have difficulty funding highway maintenance now, let alone redesigning and rebuilding our road system to accommodate driverless vehicles. (Photo: Getty Images)
Our roads and everything around them were created with people in mind. Adapting a driverless vehicle to those standards takes time, and we just aren’t equipped for a massive overhaul of the system. Reimer, in his 2018 Ted Talk, notes that it’s difficult to find funding to fix road problems, much less spend on upgrades for autonomous vehicles.
Even if we can reach a point where people fully trust their vehicles to take them safely from point A to point B, there is still a massive learning curve to overcome. People resist continuing driver’s education and balk when asked to understand new traffic laws, so it’s understandable that experts would be nervous about the prospect of turning people loose in vehicles that drive themselves.
Reimer’s belief is that advertising should be used to educate the public about how current vehicles work instead of blindly promoting the idea of self-driving cars.
Super Cruise is an advanced driver-assist system, but it falls short of making a Cadillac an autonomous car. (Photo: Cadillac)
Marketing and messaging
If you’re reading this article and are thinking “we already have self-driving cars,” you probably see where this is heading. Messaging and marketing around autonomous vehicles is a big part of the problem that the technology faces. Tesla’s Autopilot isn’t actually all that “auto,” enhancing driver safety and convenience instead of becoming the driver, and Cadillac’s Super Cruise falls into the same vein. Watching a recent YouTube clip or news report, you’d be forgiven for thinking that those technologies allow the driver to relax behind the wheel, as popular opinion seems to be that we’re on the cusp of a fully autonomous future.
As far as we’ve come with autonomy in our vehicles, we still have light years left to travel before we’re answering emails from the driver’s seat on our morning commutes. Automation can do a lot of things, but it probably won’t be the holy grail for human safety and productivity, even for something like a ridesharing scenario. Concentrating on using automation as an assistance mechanism instead of a futuristic productivity tool will help avoid disappointment and will focus consumers’ expectations on the right areas.