This weekend's accident between a self-driving Uber and a human driver in Tempe, Arizona highlights the problems that will arise as algorithm-based self-driving cars increasingly share the road with unpredictable human drivers.
The Uber vehicle, which was in self-driving mode but had a human driver at the wheel, was not at fault: It was hit by another car whose driver did not yield, a spokesperson from the Tempe Police Department told the New York Times, which led the Uber car to roll onto its side. Neither driver had serious injuries, the police spokesperson said.
The accident prompted Uber to suspend testing its self-driving cars in Tempe, Pittsburgh, and San Francisco. The company's small San Francisco fleet, which is used exclusively for mapping, was back on the roads on Monday morning. And the Tempe and Pittsburgh fleets resumed testing later in the day Monday, an Uber spokesperson told TechRepublic. The police are expected to release a full report on the crash sometime this week.
This kind of accident will become more and more common, according to Michael Ramsey, an autonomous vehicle analyst at Gartner. "Robots don't drive like humans," Ramsey said. "That's a good thing. Humans are terrible at driving, but other humans know this and adjust our driving to account for what regular people would do on the road. There are many unwritten rules of driving that humans can quickly adjust to that robots will not, and this will lead to accidents."
If all cars on the road were autonomous, accidents would decline, Ramsey said. "While they are mixed together, the inflexibility of computers may lead to accidents that wouldn't have happened before even as some other accidents are prevented," he said.
In May 2016, a Tesla driver was killed in an accident while the car was operating in its semi-autonomous Autopilot mode. A US Department of Transportation investigation did not identify any defects in design or performance of the Autopilot system. According to data released by Tesla during the investigation, Autopilot has lowered the number of crashes among its drivers by 40%.
The start of 2017 has not been easy for Uber, as TechRepublic's Hope Reese reported: Earlier this month, company president Jeff Jones left the company after only six months. In February, former Uber employee Susan Fowler Rigetti reported a culture of sexism and harassment at the company, leading to an investigation. And a "#DeleteUber" Twitter campaign launched after Uber CEO Travis Kalanick said he would take a seat on President Trump's economic advisory council. The campaign ultimately caused Kalanick to back down from the position.
Though the Tempe accident was not caused by the autonomous technology, it remains to be seen if this accident will hinder support for self-driving cars moving forward.
The 3 big takeaways for TechRepublic readers
1. A car accident between an Uber in self-driving mode and a human-manned car on the road in Tempe, Arizona demonstrates the problems that will likely arise as more autonomous cars share the road with human drivers.
2. The self-driving Uber was deemed not at fault; rather, the human driver did not yield. Neither driver was seriously injured, according to police.
3. It remains to be seen if the accident will further hurt Uber's brand, after the company was involved with a number of other incidents in recent months.
- Ditching Uber? 8 alternatives for professionals (TechRepublic)
- Uber's self-driving car to hit the streets of Pittsburgh (ZDNet)
- How data and machine learning are 'part of Uber's DNA' (TechRepublic)
- Uber slams on the brakes, stops self-driving car tests (ZDNet)
- With launch of 'Uber AI Labs,' ride-sharing giant aims to expand AI research beyond autonomous cars (TechRepublic)
Alison DeNisco Rayome has nothing to disclose. She does not hold investments in the technology companies she covers.
Alison DeNisco Rayome is a Senior Editor for TechRepublic. She covers CXO, cybersecurity, and the convergence of tech and the workplace.