Do Self-Driving Car Companies Really Care About Safety?

Or is it just a good way to market their product?

Uber vehicle after deadly crash in March 2018. Source: NTSB

A lot of big promises have been made about self-driving cars. If you listen to their boosters, the technology is poised to eliminate traffic, end car ownership, get rid of parking garages, address the carbon footprint of transportation, give us more time to work or use our phones — and those are just the ones at the top of my mind.

But biggest of all has to be the promise to save approximately 40,000 lives lost every year in vehicle crashes and collisions across the United States. The claim is always in the marketing issued by the companies working on autonomous vehicles, but can the claims of executives at Uber, Waymo, Tesla, GM, Ford, and others that their technologies will achieve this goal really be believed?

Putting aside the fact that self-driving cars are much further off than these companies led us to believe six to ten months ago — they’re all slowly admitting that the technology will take a decade, or several, to perfect — their actions beg the question of whether their commitment to saving lives is genuine, or simply a good way to promote the product they’re trying to sell to the masses (and make exorbitant profits along the way).

Waymo’s Risky Early Days

How might the development of self-driving cars take place in a way that would put safety and protecting human life as the top priority?

A lot of testing would likely happen off public roads, especially when the artificial intelligence (AI) is still struggling with basic driving functions, and when they did finally go on public roads, they would be programmed to be overly cautious to try to avoid collisions. Is that unreasonable? I don’t think so — but it seems many of the companies do. Let’s run through them.

Waymo, the Google sister company, has been held up as the leader in the space. It’s operating a very limited public service in Arizona that only launched after the company put safety drivers back in every vehicle and the CEO admitted a self-driving vehicle capable of driving in any condition would never exist. The test vehicles have recorded a number of crashes, but only recently did details of the early days of the program come to light that seriously question the company’s commitment to safety.

Anthony Levandowski, who led Google’s early efforts with self-driving cars and later moved to Uber, told the New Yorker that “[i]f it is your job to advance technology, safety cannot be your №1 concern” — a value he demonstrated at Google, under the protection of co-founder Larry Page. While heading the self-driving team, Levandowski would alter the vehicles’ software so he could take them on forbidden routes; in one incident, the Google vehicle boxed in a Toyota Camry, eventually sending the Camry off the road and into a median, and causing Levandowski’s passenger “to injure his spine so severely that he eventually required multiple surgeries.” They left the scene and never checked to see if whoever was in the Camry was okay — and that’s just one example of what could have been many similar incidents.

Uber Follows in Waymo’s Footsteps

Waymo claims safety is its top priority now, but every one of them would say that — including Uber. Yet the ride-hailing giant bought the company Levandowski founded after leaving Google, and it seems he brought the same view to Uber’s self-driving team. Leaked documents showed that former CEO Travis Kalanick and Levandowski agreed they needed a “strategy to take all the shortcuts we can” to beat their competitors; and Kalanick talked about using “cheat codes” to get ahead. Long after Levandowski was fired from Uber, the mindset he and Kalanick had infused in the team produced the deadly result that finally burst the hype bubble around self-driving cars.

On the night of March 18, 2018, one of Uber’s test vehicles struck and killed 49-year-old Elaine Herzberg as she wheeled her bike across the street. The National Transportation Safety Board’s (NTSB) report on the incident made it clear that Herzberg likely would not have been dead if it wasn’t for decisions made by the company’s self-driving team. The AI detected Herzberg six seconds before impact, but couldn’t identify what she was; it only decided it needed to stop 1.3 seconds before hitting her. But it couldn’t stop because Uber’s team had disabled emergency braking in autonomous mode to allow for a smoother ride, and hadn’t installed any way for the system to alert the safety driver to brake.

But that wasn’t all. Documents leaked to the New York Times showed that the team was under pressure to have a driverless taxi service ready to compete with Waymo by the end of 2018, yet its vehicles couldn’t go more than 13 miles (21 kms) before a driver had to intervene, compared to a disengagement rate of 5,600 miles (9,000 kms) reported by Waymo. Uber had also cut its safety drivers from two to just one in a single vehicle, which made the daunting task of trying to stay alert for hours on end even more difficult.

I don’t care what any public relations department says: these are not the actions of companies that care about public safety. Waymo and Uber are both trying to clean up their images, but only now that it’s clear self-driving vehicles are going to take a lot longer to develop. Probably the only redeeming thing that can be said for them is that they’re not selling their products to the public while misleading them about their capabilities — for that, you need the unshakeable ego of Elon Musk.

Deadly Deceit at Tesla

Tesla sells vehicles with a driver-assist system called Autopilot — already a misleading name for what it does — but Musk promotes it as far more. He hypes the system up as having the ability to drive the car, going so far as doing media interviews where he takes his hands off the wheel, even though under existing guidelines a system with the capabilities of Autopilot should require drivers to have their hands on the wheel at all times. Yet even though Musk claims to care about safety, his deceitful statements about Autopilot prove the exact opposite.

Autopilot has already killed two drivers that we know of, and has caused countless crashes. In May 2016, a Tesla Model S collided with a transport truck, killing its driver, and the NTSB cited an overreliance on Autopilot as a cause of the crash. The driver was taking his hands of the wheel and eyes off the road, and the system allow him to use it on roads it wasn’t designed for, but is it any surprise that drivers use it this way when Musk promotes Autopilot as a fully self-driving system?

Another driver was killed in March 2018 when his Tesla Model X steered him out of his highway lane and slammed into a concrete barrier. Tesla claimed that the driver ignored alerts to put his hands on the wheel, but the NTSB report clarified that no alerts had been made for more than 15 minutes before the crash occurred, yet the driver had his hands on the wheel for 34 seconds of the final minute before impact — clearly contradicting Tesla’s statement. Even more damning, the driver had been to the dealership several times to report that his vehicle had veered toward the barrier in the past and nothing had been done, until it finally took his life.

It’s worth noting that in the past few months, Tesla has stopped selling its Full Self Driving package and Musk doesn’t even seem to be setting another false deadline for Tesla to miss. But that hasn’t stopped him from continuing to hype up Autopilot beyond its actual capabilities, and that clearly shows he doesn’t actually care about the safety of his customers after two have died.

Automakers’ Wreckless History

In a piece for Wired, Aarian Marshall contrasted the self-driving efforts of tech companies and automakers: “If the Silicon Valley motto is ‘move fast and break things’, Detroit’s seems to be ‘move below the speed limit and ensure you don’t kill anyone’.” This places the automakers in a better light — as the ones who actually care about safety — but do they really?

The automakers have embraced the safety narrative promoted by tech companies, but have a long history of opposing concrete improvements to automotive safety. Columbia University professor Steven Cohen gave a quick overview of their efforts in a piece for HuffPost:

Over half a century ago, the American auto industry fought against seat belts. In 1970, they fought against the requirement that they install catalytic converters to reduce air emissions. They have opposed airbags and other safety standards as well. Today, they continue to fight against reducing emissions and improving fuel economy.

Given their history, it’s reasonable to approach the auto industry’s claim to care about safety with skepticism. In late 2018, the companies developing autonomous vehicles were pushing federal lawmakers to pass AV START, a law that would allow them to test their technologies on public roads without having to meet existing safety standards or provide public reports on their activities. That doesn’t sound like the actions of an industry that holds safety as its top priority.

Move Slow and Save Lives

The truth is that we don’t need to wait a decade or however long it will take to have reasonably capable self-driving cars to start reducing the number of automobile deaths; we already have all the technology we need to save tens of thousands of lives every year.

The United States is an outlier among other developed countries on driving deaths because other countries have moved forward with policies to make driving safer: lower speed limits, stricter limitations on drunk driving, more cameras, higher rates of seatbelt use, and more. But they’ve also been redesigning roadways under the framework of “Vision Zero” to build safety into the infrastructure itself.

And that’s not all. The best way to reduce driver deaths is for fewer people to drive cars, and instead switch to transit, cycling, and walking in order to get where they need to go. Cities around the world are increasingly prioritizing transit and cycling in their city centers, making major investments in new infrastructure, and embracing the proliferation of dockless bikes and scooters to introduce residents to active forms of transportation.

Waymo, Uber, Tesla, GM, Ford, and others claim that safety and reducing automobile deaths is their top priority, but the actions of these companies contradict their stated intent. If they really cared about safety, they wouldn’t be putting unproven technologies on public roads, misleading the public about the capabilities of their products, and fighting government regulations designed to save lives. I hope these companies do put safety first, but so far it looks more like a marketing slogan than a real commitment.