Self-Driving Car Makers Prepare to Blame “Jaywalkers”

AV crash

Who’s to blame when an autonomous car hits a pedestrian? Car companies are already trying to force the question for their own benefit.

In the months after an Arizona pedestrian was killed by a driverless car, tech companies developing the technology are trying to shift blame to those on foot, as it becomes increasingly clear that self-driving cars are having trouble detecting pedestrians, reports Jeremy Kahn at Bloomberg:

Driverless proponents … say there’s one surefire shortcut to getting self-driving cars on the streets sooner: persuade pedestrians to behave less erratically. If they use crosswalks, where there are contextual clues—pavement markings and stop lights—the software is more likely to identify them.

“What we tell people is, ‘Please be lawful and please be considerate,’” Andrew Ng, a machine learning researcher whose venture fund invests in driverless startups, told Bloomberg.

In other words, the paper concluded, “no jaywalking.”

Elaine Herzberg was, technically, crossing illegally when she became the first pedestrian killed by a self-driving car this spring — a crash that revealed major problems with autonomous tech. The Uber Volvo SUV that hit her as she walked her bike in a pedestrian-heavy area had a hard time identifying her, plus the car was programmed not to brake if it believed it had detected “false positive.” The National Transportation Safety Board has not yet issued its final report on that crash — including who would be at fault.

But there is more than enough evidence from NTSB’s preliminary report to show that Uber made a lot of dangerous mistakes. For example, the backup driver was watching a television show on her phone at the time of the crash and Uber had programmed the car to avoid braking in part because it made an active decision to reprioritize safety in its rush to get to market.

But jaywalking is something that people do — and, indeed until fairly recently, did with impunity. Laws to criminalize jaywalking were initially promoted by car companies about a century ago to shift blame for traffic injuries and deaths to the pedestrians they were killing.

“If you ask people today what a street is for, they will say cars,” Peter Norton, an assistant professor at the University of Virginia and the author of Fighting Traffic: The Dawn of the Motor Age in the American City, told CityLab. “That’s practically the opposite of what they would have said 100 years ago.”

And now we have Andrew Ng and his high-tech cohort that wants to go further than merely summonsing pedestrians — they want to turn pedestrians into robots so that the robot cars can avoid them. (As a point of information: Ng’s quote itself suggests he doesn’t understand crosswalk laws. Every intersection is technically an unmarked crosswalk — a legal crossing zone — even without “contextual clues” like stripes and traffic lights that could help alert computer systems.)

Beyond that, self-driving cars are billed as a major safety breakthrough to rid us of the single biggest adverse effect of human drivers: how frequently and violently they kill their fellow road users. Driverless proponents have argued they shouldn’t be subject to existing safety regulations because anything that delays the widespread adoption of autonomous vehicles could cost lives.

So it’s disturbing to see promoters placing responsibly for safety now on external actors, especially those most vulnerable to their shortcomings.

A potentially larger question — who will be liable in crashes between driverless cars and pedestrians — remains largely open.

The Governor’s Highway Safety Association recently discussed the situation in a State-Farm Insurance-sponsored report. The organization laid out a scenario in which a pedestrian signals to a car that he is planning to cross mid-block, but is struck anyway. Who’s at fault in these situations, GHSA wonders?

It is sad that this is even a question open for debate. “Jaywalking” shouldn’t be punishable by death. And pedestrians should not be re-educated into robotic machines that move in predictable ways to meet the demands of programmers of robotic cars. It’s supposed to be the other way around.

But not to the GHSA, which suggests that “new public outreach or even enforcement efforts” might be needed to make sure pedestrians stay in line.

More disturbingly, other tech industry insiders are eager to unleash their programming against “repeat offender” jaywalkers who mess with their precious code.

“Pedestrians and pranksters, knowing that the cars are programmed to yield to any in their path, could bring traffic to a halt,” CNN Tech reported, apparently giving voice to the techies. “Outfitting the cars with facial recognition technology could help identify violators, but that raises its own tricky issues.”

For now, the Twittersphere is still treating this as a joke …

But the future has the potential to be scary — and not just because there’s no one behind the wheel of that 2,500-pound metal cage speeding down the roadway. What’s truly scary is that driverless car makers seem to want to shift the blame for crashes onto pedestrians who “misbehave” or, in other words, resist the coders’ re-education campaign.

And that debate is not about safety, but about accountability. If humans have to alter their behavior to accommodate the machines they create, we’re one step closer to a sci-fi dystopia.

117 thoughts on Self-Driving Car Makers Prepare to Blame “Jaywalkers”

  1. You don’t think people will adapt, knowing that the new technology may not be able to see them as well? People adapted to typing with their thumbs on the older (early ’00s) cell phones, they can adapt to the way this technology will behave.

  2. I think you’re missing the point. The debate isn’t about whether there is a potential safety benefit or whether safety is a worthy goal. It’s about how we are going to achieve that goal and, more importantly, whose interests we should prioritize in the pursuit of that goal.

    Early 20th century car companies were able to prioritize their interest in selling cars over the interests of people living in cities by reframing the debate about pedestrian behavior rather than about the practicality or desirability of cars as a primary means of urban transport. Combined with the fact that auto lobbies successfully integrated car-oriented design preferences into building codes preferred by FHA loans and the result was a built environment that, in the vast majority of cases privileges cars over every other means of transportation and in most cases in America all but necessitates cars as your only way of getting around. Thus, a clear windfall for car manufacturers to the detriment of everyone else.

    We are therefore right to be suspicious of the AV industry who now seem to be parroting the same line. They would much prefer a world that restricts pedestrian activity and privileges AVs over walking. Picture, on the one hand, your city with pedestrian fences lining every meter of sidewalk except at “designated crosswalks” and the streets filled with zooming single occupancy AVs. Now picture, on the other hand, that same city without the fences and with freedom of pedestrian movement and the occasional self-driving bus passing through (filled with 40-60 passengers) and a complete ban on single-occupancy vehicles in the city center. Assume that in both cases we have achieved both safety and mobility goals. The only difference is that in the former the AV companies have managed to sell a lot more cars / private subscriptions. Which of those two futures do you prefer? And which one do you think Uber prefers?

  3. “I’ve never understood why we privatized cameras. Seems like a big giveaway to well-connected individuals” Read that again, slowly. Now read it again. It answers itself…..

  4. I’ve read that each state can legislatively decide whether cameras can be used to enforce traffic violations. What a mess that will be wihen the cameras are in the cars. Also that England uses them on all their motorways at fixed intervals and issues speeding tickets based on the time between them. Now that could almost pay for their state installation on “free”ways, but those are federal usually.

  5. Even if they succeed in blaming pedestrians, that still won’t help self-driving cars avoid hitting deer, cattle, moose, and other random stuff. A deer strike can destroy your car, and can also cause (human) injury. Hitting a moose is like hitting a stationary SUV. Last week I narrowly avoided hitting a whole truckload of stuff that had fallen into the road… at 70mph.

  6. Unfortunately that’s a pipe dream with our current infrastructure. A much higher percentage of people in the United States drives partially because of laxer restrictions on driving, but also because our cities are overwhelmingly built for driving unlike much of Europe.

    This is further compounded by the reality that a large percentage of people will drive without a license precisely because living in this country simply doesn’t work without it for most people.

  7. I like pipedreams better than dystopian nightmares. There’s lots of methods to nudge people towards calmer driving. I’ve been preaching education for forty years now to retrain road users, but engineering and enforcement get the majority of what little funding there is for road safety, most monies are for speeding traffic up, not slowing them down. We are different than Europe, but many of their innovations would work here, if translated to local conditions.

  8. It won’t work from the direction of education. Europe can do what it does because driving isn’t as baked into it’s infrastructure. Outside a very small set of large cities driving will be done with or without a license.

  9. you’re right that people will drive with or without licenses, but when undocumented immigrants were allowed to get licenses in California, millions of them did. It’s nerve wracking to drive without one, even if you believe you’re a sovereign citizen. Licensing doesn’t necessarily make drivers safer, people driving without them are probably more likely to follow traffic laws, if they know what they are. Of course education can work, we just haven’t put the money or effort into multi-modal training and reinforcement.

  10. You are quite missing the point. The principle problem is that driving is a necessity in most places in this country. Until you create a world in which it is generally viable for people to opt-out of driving you will train the masses to achieve European levels of safety in vain.

    The question isn’t whether someone does or doesn’t have a license. That merely proves that people WILL drive with or without the license. They do so as a rule because the alternatives are untenable.

    Not everyone is mentally or physically capable of driving at the level of safety you are looking for. Against that truth all efforts crash like waves against the shore. Brave, loud, and powerful, but ultimately the beach remains eternal.

  11. I like the metaphor, and I agree that many people aren’t capable of driving or cycling safely now. I have moments of inattention driving, even realizing it’s the riskiest activity I’m likely to do on a daily basis. Yet, considering how many millions miles are traveled, for the most part we have learned to operate in an incredibly complex and hazardous system, without particularly rigorous or mandatory training.
    We can greatly decrease collisions between cars and cyclists/ pedestrians without building a Dutch style infrastructure, if we concentrated on slowing motor vehicles down in urban areas and suburban neighborhoods. Motorists (and cyclists to some extent) can learn to share the road instead of owning it. Education alone isn’t enough, but we haven’t done enough of it to know how much more effective it could be.
    Waves of mass education from many sources, continuous training and peer pressure to drive, or cycle or walk as if they were part of a system, with clearly defined roles and rules is probably the most affordable way to make it safer and more pleasant to use these forms of transportation.

  12. You miss my point entirely once again. It isn’t a matter of now. They cannot be taught at all. They haven’t achieved it in Europe. It is simply that many less people drive.

  13. I haven’t been in Europe for forty years, but from how many died in the recent bridge collapse it looks like just as many drive now as they did in the 80’s and that was a lot. They’ve gotten much safer per mile driven,while Americans have only improved safety a fraction of their rates. People are teachable, but it’s slow and frustrating without the resources needed.

  14. Actually, it is true, besides the state laws, most local municipalities replicate a version of this Long Beach code and heavily enforce. I’ve been stopped myself for crossing in the middle of street by police and got off with a warning only because I had a chicago drivers license and was visiting. They wrote me up, gave me a free warning ticket, and said hte next one would cost me since I’d been warned…

    Municipal Code Volume 1, Title 10 Vehicles and Traffic, Chapter 10.58 Pedestrians
    § 10.58.020. Crossing without crosswalk
    “No pedestrian shall cross a roadway, other than by a crosswalk, in the Central Traffic District, or in any business district, except at intersections where pedestrian traffic is controlled by a scramble-system automatic signal.”

  15. It’s not punishment. It’s a consequence. Punishment implies that the death is justified or deserved. A pedestrian who breaks the law doesn’t deserve to die. They deserve to be ticketed for their offense. Death by being stuck by a vehicle is the consequence in this case. It doesn’t always result in death but it is a known risk. The consequence is situational and the responsibility of all parties involved. To fix a problem, it is important to look at the whole picture and not just one side of it. Everyone involved, working together to improve a scenario like this in the future, is what’s most productive.

Leave a Reply

Your email address will not be published. Required fields are marked *


How the Self-Driving Car Could Spell the End of Parking Craters

Here’s the rosy scenario of a future where cars drive themselves: Instead of owning cars, people will summon autonomous vehicles, hop in, and head to their destination. With fewer cars to be stored, parking lots and garages will give way to development, eventually bringing down the cost of housing in tight markets through increased supply. […]