The Driverless Commute: Urban planning nightmares with AV deployment; AV industry’s diversity and inclusion crisis; and how bogus satellite data could hack an AVs operation.

Print Friendly, PDF & Email

1. When AVs are king

New York City

Cities were once highly compact, walkable places that blended residences and workplaces and where people commanded primacy. Then the car came along.

Now, the modern American city, sprawling and traffic-plagued, is an ecosystem in complete service to cars. But what if AV deployment invites an even deeper calcification of the cars-first mentality in city centers?

Just imagine sidewalk gates. That was the whacky idea floated by one unnamed “automotive industry official” in a recent New York Times article:

“In New York, the unwritten rule is plain: Cross the street whenever and wherever — just don’t get hit. It’s a practice that separates New Yorkers from tourists, who innocently wait at the corner for the walk symbol. But if pedestrians know they’ll never be run over, jaywalking could explode, grinding traffic to a halt.

“One solution, suggested by an automotive industry official, is gates at each corner, which would periodically open to allow pedestrians to cross.

“That prospect seems as likely as never-late subways. But it’s an example of the thinking by those who worry about planning for the future.”

Autonomous vehicles will deploy in the real world, where pedestrians jaywalk and motorists drive aggressively. Trying to engineer solutions—if you want to call it that—to change ingrained social structures is a fruitless endeavor. Self-driving cars need to be able to withstand the chaos of real-world roads, not tech-bros’ vision of utopia.

The Driverless Commute is provided by Dentons’ global Autonomous Vehicles team. If you believe a colleague or associate would benefit from this service, please share this link so they may subscribe.

2. Gender gaps and racial blind spots: how the AV industry’s lack of diversity is creating new problems for deployment

The AV enthusiasm and trust gap between men and women has surged to an almost a 20-point difference in the latest research by AAA.

A majority of Americans remain wary of self-driving cars, but skepticism of the technology among women is especially pronounced.

  • Seventy-nine percent of women are afraid to ride in an autonomous vehicle, while only 62 percent of men told pollsters the same.
  • Similarly, only 14 percent of women said they were comfortable with the notion of full autonomy, while 30 percent of men said they were.

Those results track with a 2016 AAA survey but stand in contrast to recent efforts by makers of self-driving cars, most notably Waymo, to demonstrate the technology’s reliability after a series of notable industry stumbles. So what explains the durability of women’s distrust?

New York University Professor Meredith Broussard, an expert on AI implementation bias, told Axios the lack of representation and consideration of women in the engineering of these cars could be driving their suspicion. (By our count, the only prominent AV firm led by a woman is Zoox, whose female CEO only ascended earlier this year.)

The industry’s diversity and inclusion problem isn’t just a problem for women, but could also pose safety risks for people of color.

Researchers at the Georgia Institute of Technology tested the accuracy of object detection systems (not unlike those used in driverless cars) in positively identifying pedestrians of varying skin color and found a uniformly poor success rate in spotting people of color compared to those with lighter skin tones.

  • The researchers posited that the technology wasn’t inherently capable of recognizing people of color.
  • Instead, it was the dataset that trained the system that was to blame: the computer was trained to recognize pedestrians using a data set of predominantly white objects, so it was better equipped to recognize these faces.

If the data on which artificial intelligence is trained skews towards the biases of its engineers (like white men), those biases will be reflected in the computer.

3. The Auto(nomous) Bahn