<img height="1" width="1" src="https://www.facebook.com/tr?id=769125799912420&amp;ev=PageView &amp;noscript=1">
ADVERTISEMENT
About Our People Legal Stuff

Safety advocate: Deadly self-driving Uber crash 'exactly what we've been afraid of'

0

Investigators have not uncovered evidence indicating that a self-driving Uber vehicle was at fault in an accident that killed a pedestrian in Tempe, Arizona Sunday, but the incident may still trigger a public reassessment of the fledgling technology as manufacturers attempt to step up street testing.

The self-driving Volvo SUV struck 49-year-old Elaine Herzberg when she stepped off of a center median about 100 feet from a crosswalk. According to police, the car was moving at 38 miles per hour and it did not attempt to brake.

“A crash like this showcases the risk of making humans the guinea pigs in self-driving vehicles tests,” said Jason Levine, executive director of the Center for Auto Safety. “Clearly, driverless vehicles are not yet ready to be tested in uncontrolled environments and on city streets.”

The accident is still under investigation but Tempe Police Chief Sylvia Moir told the San Francisco Chronicle a review of videos from the car’s two cameras indicated it could not have stopped in time, no matter who was in control.

“It’s very clear it would have been difficult to avoid this collision in any kind of mode (autonomous or human-driven) based on how she came from the shadows right into the roadway,” Moir said.

As Uber and many other companies put self-driving cars on the road, Arizona has emerged as a popular destination after Gov. Doug Ducey decided to welcome them “with open arms and wide open roads” in 2016. More than a dozen California cities are also hosting test runs of vehicles, as are Boston, Detroit, Atlanta, Miami and several others.

Proponents of the technology argue that self-driving cars could prevent the approximately 90 percent of motor vehicle accidents that are caused by some form of human error. And while some critics do not dispute that, they do maintain the technology is too new and too unreliable to be operating with minimal oversight on city streets across the country.

ADVERTISEMENT

Cathy Chase, president of Advocates for Highway and Auto Safety (AHAS), stressed that a National Transportation Safety Board (NTSB) investigation will be the final word on the cause of the Tempe crash and initial assessments are not always correct.

“People are quickly moving toward that assessment but I still would like to hear what the NTSB has to say,” Chase said.

The inability of autonomous cars to adapt to sudden erratic behavior by others does not fill her with faith in the technology because pedestrians, cyclists, and other drivers sometimes are erratic.

“Isn’t the whole point that they’re supposed to be safer than human drivers?” she asked.

Regardless of who is determined to be at fault, the wave of headlines about a self-driving car killing a pedestrian could undercut confidence in an emerging technology that the public was already reticent to embrace.

“It’s too early to tell how this will play out, but certainly from the perspective of how we think about risk, I suspect there are going to be issues beyond technically who is to blame,” said Andrew Maynard, director of the Risk Innovation Lab at Arizona State University.

Amid existing uncertainty about autonomous vehicle technology and its development, the accident shines a spotlight on broader issues and increases the imperative for companies to earn the public’s trust.

“There’s some damage to the public confidence,” said Bart Selman, a professor of computer science at Cornell University who studies artificial intelligence. “However I think the lasting effect will be minimal because in the bigger picture people realize these cars will actually save lives in the long run.”

ADVERTISEMENT

He cautioned against drawing a conclusion that these cars are unsafe based on one accident.

“The cars, under good driving conditions with good visibility, clear roads, they are already as safe or safer than human drivers,” he said.

Whatever the facts are, one group that supports automation suggested lessons learned from the crash will make future cars even safer.

“This is a tragic and unfortunate incident,” said Brian Wynne, president and CEO of the Association for Unmanned Vehicle Systems International. “We understand that authorities are investigating and it’s important we understand what happened to prevent incidents such as this from occurring in the future. Automated vehicles hold tremendous promise to usher in a new age in mobility and eventually make our streets safer once the technology fully matures.”

In the wake of the accident, auto safety advocates have raised a number of questions about how road-ready these vehicles are and how prepared the government is for the regulatory issues they will create.

“[The National Highway Traffic Safety Administration’s] completely voluntary autonomous vehicle guidelines have unfortunately put the manufacturers into the driver's seat when it comes to the roll-out and testing of self-driving cars,” Levine said. “Not only has the rush to get self-driving vehicles on the road resulted in several crashes and now a pedestrian death, but it has put the future of these vehicles at risk.”

While it may seem that self-driving cars have gone from theory to public testing with human passengers quite fast, some experts say the speed is commensurate with the rate at which the technology has advanced over the last decade.

“If you look at the progress that has been made, the capabilities of the cars at this point, it seems to be the right pace,” Selman said.

According to Maynard, who lives in Tempe and sees self-driving vehicles on the streets every day, a testing period like this is necessary for any new technology.

ADVERTISEMENT

“I think what this particular incident does is it raises a flag that says, maybe we’ve got to move faster in terms of ensuring safety instead of just improving the technology,” he said.

Prior to the crash, the American public had been slowly coming around to acceptance of driverless cars. According to a poll released by AAA in January, 63 percent of Americans were afraid of riding in a self-driving vehicle, down 15 points from the same time in 2017.

A poll commissioned by AHAS around the same time identified a number of areas where a vast majority of the public wanted stricter federal standards. These included minimum standards for new features (73 percent support), cybersecurity standards (81 percent), standards governing human ability to take control (84 percent), and minimum performance requirements for operating computers (80 percent). In addition, 63 percent opposed granting autonomous vehicles mass exemptions from existing safety standards.

Chase and other safety experts fear legislation under consideration in Congress now will not sufficiently address these issues. The AV START Act awaiting a vote in the Senate and a similar bill that already passed in the House would allow developers to test and market self-driving vehicles before federal regulations are finalized, and they would prohibit states from imposing tougher regulations than the federal government.

“A crash like this is exactly what we’ve been afraid of and exactly why we’ve been pushing for improvements to the AV START Act,” Chase said.

Experts and advocates agree that safe, reliable autonomous vehicles will provide great benefits for the public. The big question is what to do until they are safe and reliable enough.

“Advocates are not opposed to the concept of autonomous vehicles,” Shaun Kildare, research director for AHAS, said. “We know it has tremendous promise, but it has to be done in a safe manner.”

Observing that a Senate committee was holding a hearing Tuesday on defective Takata airbags that killed at least 20 people, Chase said the auto industry has proven it cannot be counted on to police itself. AHAS has submitted a letter to the committee arguing for more stringent standards on autonomous vehicles to avoid similarly disastrous results.

ADVERTISEMENT

“We should not allow history to repeat itself especially one replete with industry malfeasance affecting millions of consumers and needlessly causing deaths and injuries,” the letter states. “Now the same industry is asking the public and the government to ‘trust them’ as they develop and deploy new driverless car technology.”

Levine also faulted the lax approach to regulation so far, calling for a much more methodical and controlled testing process before the vehicles are set loose on city streets.

“There is no doubt that self-driving vehicles have the potential to save many lives--but we need to be careful in how we get there,” he said. “The race to be first to get these vehicles on the streets is less important than getting the technology right.”

In the short term, Maynard suggested access to these vehicles could increase mobility for the elderly and disabled and make teens and new drivers safer. In the long term, cities of the future could integrate fully autonomous transportation systems.

“This is almost like the holy grail of where we go with autonomous vehicles,” he said. “The challenge is we won’t see that if we don’t develop the technology responsibly now.”

GAO Automated Vehicles by Stephen Loiaconi on Scribd

Some researchers warn that waiting until the technology is perfected could result in hundreds of thousands of needless deaths.

ADVERTISEMENT

A study released by the RAND Corporation last November estimated that if self-driving cars are made available to the public once they are deemed to be slightly safer than human drivers, 1.1 million lives could be saved over the next 80 years. If they are not introduced into the market until they are nearly perfect, they might only save half as many.

At the time, researcher Nidhi Kalra predicted that a high-profile crash could set the introduction of the technology back decades.

“A major backlash against a crash caused by even a relatively safe autonomous vehicle could grind the industry to a halt—resulting in potentially the greatest loss of life over time,” Kalra said. “The right answer is probably somewhere in between introducing cars that are just better than average and waiting for them to be nearly perfect.”

Safety advocates will be watching closely as Uber and other companies respond to this accident, and as state and federal lawmakers decide whether to change their posture toward regulation of autonomous vehicle testing.

“If manufacturers want to ensure consumer acceptance of self-driving cars, they need to prove to consumers that they are safe,” Levine said.

Barring a backlash, experts say self-driving cars could be common sights across the country within about five years. They will likely roll out first through transportation and ridesharing services because the technology will initially be very expensive for individual drivers.

The key to preventing progress from slowing down or reversing, according to Maynard, is for the industry to engage with various stakeholders in communities and address the concerns the public has.


“You have to make the consumers part of the process of innovation,” he said.

ADVERTISEMENT

Selman said the crash should serve as a reminder that accidents involving autonomous cars will still happen. He expects the circumstances surrounding those accidents will be different and somewhat unusual because robots react to situations differently than humans.

“It would be too much to expect no accidents at all with self-driving cars,” he said.

Comments
Read Comments
Comments
ADVERTISEMENT
Facebook Twitter Instagram Pinterest Linked In List Menu Enlarge Gallery Info Menu Close Angle Down Angle Up Angle Left Angle Right Grid Grid Play Align Left Search Youtube Mail Mail Angle Down Bookmark