Who do Millennials Blame for the Uber Self-Driving Car Accident?

by Imran Ali and Tal Bar-Or

National Transportation Safety Board investigators examine the Uber self-driving car that was involved in a fatal accident on March 18, 2018. Credit: Reuters

As the first fatality caused by a self-driving car in the U.S., the accident that killed a pedestrian in Arizona on March 18 was a somber watershed moment for the burgeoning autonomous vehicle industry.

Spurred by this tragic incident, we conducted a survey of 100+ millennials to better understand their perceptions of self-driving car safety. Our goal was to develop actionable steps for self-driving car companies to take to continue building trust with the public.

We found that millennials remain enthusiastic about this technology, with over 90% willing to ride in a self-driving car. Additionally, we found:

  1. Nearly 1 in 3 blame the pedestrian for the Uber accident.
  2. Millennials who are are more familiar with self-driving car technology are more likely to blame the technology or Uber.
  3. Millennials largely believe that the technology is safe for its passengers, but are more wary for pedestrian/cyclist safety.
  4. Watching the dashcam footage of the accident produced no significant change in the perception of self-driving car safety.

In order to continue building trust with this key demographic, we recommend self-driving car companies:

  1. Partner with the public sector to invest in infrastructure improvements.
  2. Avoid over-promising the capabilities of their technology.
  3. Provide back-up drivers with additional safety resources.

Our Survey Demographics and Insights

We received 135 responses, of which we kept 93 responses based on our screening criteria*.

Our survey-respondents are a target demographic for early adoption of self-driving car technology. Respondents were millennials between the ages of 25–35 who mostly live in major American metropolitan areas, where these cars will be launching first. Most of our respondents drive a car no more than a couple times a year, but use a rideshare/taxi service at least several times per week. 81% of survey-respondents walk alongside a street for over 10 minutes at least several times per week.

We asked our survey-respondents their thoughts on the safety of self-driving cars before and after asking them to watch the dashcam footage of the moments leading up to the Uber collision. We found surprising and interesting insights including:

Nearly 1 in 3 people blame the pedestrian for the accident. Despite many technology experts blaming Uber’s processes or technology for causing the accident, 29% of our respondents listed the pedestrian as “primarily at fault for the self-driving Uber accident.”

When blaming the pedestrian, respondents commented:
“It was dark and hard to see the pedestrian, it seems that he [sic] crossed the street without looking and didn’t wear any glowing gear.”
“Pedestrians need to look while crossing roads and assume cars/drivers do not see them.”
“Walking in unlit area away from crosswalk. An attentive driver would likely not have been able to stop.”

In the months leading up to this collision, Uber took many aggressive steps to launch their vehicles more quickly. Uber went from two backup drivers to one (most of the competition still has two during testing). They had turned off the Volvo’s safety equipment to test the software. And finally, they only had one LiDAR sensor on top of the vehicle (most of the competition has sensors on the sides as well). In fact, after the accident, Waymo (Google’s self-driving car division) CEO John Krafcik claimed that their technology would have spotted the pedestrian and avoided this tragedy.

The more people are familiar with self-driving car technology, the more likely they are to blame the technology or company for the Uber accident. If they are less familiar with the technology, they tend to blame the individual people, such as the pedestrian or backup driver, for causing the accident.

People that are familiar with the technology may expect LiDAR and RADAR sensors to detect objects in the nighttime even when humans cannot. Those who do not understand the technology as well may think self-driving cars detect objects visually like human drivers. As a result, they do not expect the car to have seen the pedestrian crossing in the dark.

Millennials largely believe that self-driving car technology is safe, but are more wary for pedestrian/cyclist safety than passenger safety. 81% of respondents believe that self-driving cars are not dangerous for passengers, however only 62% believe that they are not dangerous for pedestrians/cyclists. We worry that there may be a backlash to this technology if this gap grows larger.

Watching the dashcam footage of the accident produced no statistically significant change in the perception of self-driving car safety for either passengers or pedestrians/cyclists. This either means that this footage did not provide viewers with any new information to change their perspective, or that millennials understand that this technology is still very nascent and they have a higher willingness to accept one-off accidents in the near-term. While this accident was certainly tragic, there may unfortunately be other fatal accidents in the future, which may shape millennial’s safety perceptions

*Screening criteria: 25 – 35 age range, watched full length of the Uber accident dashcam footage, & completed full survey

What can self-driving car companies do?

It is crucial for self-driving companies to retain the trust of the public in order to successfully launch these cars. Currently, the good news for self-driving car companies is that over 90% are willing to ride in a self-driving car, with 43% ready to ride in one today.

However, the way self-driving car companies react to these incidents and proactively work with regulators to address the issues will likely shape millennial perceptions moving forward. We have outlined some initial steps that these companies can take in order to continue building trust:

Self-driving car companies should partner with local governments to invest in infrastructure improvements. More millennials think pedestrians/cyclists on the road are in greater risk than passengers in a self-driving vehicle. To improve both perceived and actual safety for pedestrians and cyclists, companies should partner with cities to understand the unique road environment of each city and identify areas where safety can be significantly improved. For example, cities should ensure dedicated car, bike, and pedestrian lanes and spaces, with clear demarcations and physical barriers to separate them. These companies should publicize their efforts, so that the general public understands the private and public sector are working together, which will likely improve the perceptions of pedestrian/cyclist safety in areas where they are testing.

Self-driving car companies should avoid over-promising the capabilities of their technology. Since our respondents with higher levels of familiarity with self-driving technology placed more blame on the technology over the pedestrian, companies should be very careful and precise with how they portray their technology’s abilities, in order to maintain the trust of those who know more about the technology. Since 50% of respondents will either wait for a widespread adoption or a recommendation from a friend/family member before they ride in a self-driving car, trust in the technology is crucial for effective deployment and adoption.

While we applaud Waymo’s efforts to explain self-driving technology in initiatives such as their Waymo 360° Experience: A Fully Self-Driving Journey video, we also encourage self-driving car companies clearly establish the limits of the technology in order to more effectively build the public’s trust. We understand the challenges associated with this type of communication, given the battle for talent and funding, but we think this will positively benefit the field in the long-run.

Self-driving car companies should provide back-up drivers with additional safety resources. Self driving car companies continue to test in increasingly demanding environments. Although Waymo has achieved the first driverless car rides without backup drivers in Arizona last fall, back-up drivers are likely needed for most terrain in the foreseeable future. Waymo and Cruise Automation, often considered the two leaders in the space, continued to have driver disengagements (which is when a driver has to intervene during self-driving mode) in 2017. Back-up drivers in Cruise vehicles on average disengaged self-driving mode once every 1,254 miles, and Waymo back-up drivers disengaged on average once every 5,596 miles.

Despite our respondents placing the lowest amount of blame (14%) on the back-up driver, we worry that this number could increase if the number of accidents increase. We encourage self-driving car companies push for regulations that require two back-up drivers until no disengagements are experienced for an extended period of time in the area being tested or develop internal procedures requiring this. We recommend self-driving car companies invest in implementing safety technologies that monitor back-up driver attention. Nauto’s bidirectional cameras are an effective way to monitor and nudge back-up drivers when their attention begins to drift. Additionally, Cadillac has developed a patented semi-self driving technology called Super Cruise with a camera monitoring the driver, which disengages self-driving mode if the camera detects the driver’s eyes not on the road.

We believe that developing self-driving technology can be a noble pursuit, which can save lives and increase the quality of life for people who can’t drive, such as senior citizens and disabled individuals. While legal experts have pointed to product-liability law as a way for determining the legal fault of a self-driving car accident, it will take time for legal norms to develop in this field, and safety perceptions will likely take hold first.

President Barack Obama discussed self-driving technology in September 2016 by commenting, “The quickest way to slam the brakes on innovation is for the public to lose confidence in the safety of new technologies.”

Many millennials continue to accept self-driving technology as being safe, even in light of this tragic accident. However, if these companies appear to engage in malfeasance or risky behavior, this trust may vanish quickly.

Imran Ali and Tal Bar-Or are MBA students at Wharton and co-founders of Whartonomous Vehicles: Future Mobility Club, a club dedicated to engaging Wharton MBA students interested in tomorrow’s mobility ecosystem.

This is our first piece in an effort to learn more about perceptions of self-driving cars. If you’re interested in learning more or have any feedback, please drop us a line at imranali@wharton.upenn.edu and btal@wharton.upenn.edu