Human Control of Self-Driving Cars on Policymakers’ Radar
By Michaela Ross, BLOOMBERG BNA
January 11, 2018
Should human beings be able to take the wheel from self-driving cars?
Policymakers nationwide are debating that question: whether autonomous vehicle companies should be required to have features that allow humans to take control of self-driving vehicles in case they fail.
The issue has surfaced in Congress and at the state level as car and tech companies, like General Motors Corp.'s Cruise LLC and Alphabet Inc.'s Waymo, race to roll out self-driving cars. Some companies are already developing such technology. Lawmakers’ debates over safety could hasten development of the technology.
Sen. Diane Feinstein (D-Calif.) and Sen. Richard Blumenthal (D-Conn.) have stalled federal legislation that would speed up the rollout of autonomous vehicles, because they worry the new technology is not safe enough for public roads. Blumenthal said in December that he’s seeking a provision to address the issue.
“As it stands this bill does not include enough protections to keep drivers, passengers, and pedestrians safe, including robust and proven fallback mechanisms that will take control when artificial intelligence fails,” Blumenthal told Bloomberg Law Jan. 8 in a statement. He said he was having “productive conversations” with the bill’s co-sponsors to find a legislative solution.
The California Department of Motor Vehicles is aiming to finalize rules in early 2018 that would require self-driving vehicle testers to have remote human operators who can maneuver driverless vehicles during testing, a DMV spokesperson said.
Florida and Nevada have statutes that could prompt self-driving vehicle operators to hire remote human controllers. Their laws allow for autonomous vehicles to operate without a human driver inside, but also require human operators—remotely or not—to be in position to take control in case self-driving technology goes on the blink.
Driving Via Real-Time Video
Some companies, such as Nissan Motor Co., Aptiv Plc and Phantom Auto, are working on remote operator technology that could assuage regulator concerns.
States “want AVs deployed in as safe a manner as possible, and they’re starting to see that remote operation technology is a viable way to fulfill that goal,” Elliot Katz, co-founder of Phantom Auto, told Bloomberg Law.
Phantom, which is demoing its technology at the CES technology trade show in Las Vegas the week of Jan. 8,, says its system allows human operators to remotely access a driverless vehicle, see real-time video of the environment around it, and navigate it in low speed, complex conditions, such as road construction.
Nissan is developing the technology for its cars and believes that human support is necessary for complex situations, Maarten Sierhuis, director of the Nissan Research Center told Bloomberg Law. Its technology would allow a remote operator to “paint” a virtual path on a map for the car to follow and navigate it onto public roads.
Remote operator technology has risks that policymakers should consider, says John Simpson, Consumer Watchdog’s privacy project director. Such operators would be difficult to deploy at mass scale if driverless cars proliferate, Simpson said. The technology would also reintroduce human error, he said.
“You’d lull people into thinking the technology is safer than it actually is, and that’s a terrible thing” he told Bloomberg Law.
A system in the car that communicates with the remote operator could be vulnerable to hackers, Norma Krayem, co-chair of Holland & Knight LLP’s cybersecurity and privacy team, told Bloomberg Law.
“Any remote connections or potential control issues must include robust cybersecurity, or it simply adds cyber risk that will need to be managed,” Krayem said.
To contact the reporter on this story: Michaela Ross in Washington at [email protected]
To contact the editor responsible for this story: Roger Yu at [email protected]