DRIVING THE
FUTURE
The Santa Clara Law Review Symposium Explores the Promises
and Challenges of Autonomous Cars.
How would you feel if you were driving along a road and you saw two tons of steel without a driver barreling toward you at 60 miles an hour? Terrified?
Those envisioning the cars of tomorrow say you should feel relieved.
“Ninety-five percent of automobile accidents are caused at least in part by driver error,” said Sven Beiker, executive director of the Center for Automotive Research at Stanford (CARS). “If we eliminate the driver, we eliminate a lot of risks.”
Beiker was one of 25 professionals in law, engineering, policy, and insurance from around the country who spoke at the 2012 Santa Clara Law Review Symposium, “Driving the Future: The Legal Implications of Autonomous Vehicles,” co-sponsored by the High Tech Law Institute and State Farm Insurance, on January 20. Over nine hours, speakers and attendees discussed new technologies for creating cars that drive themselves, and the challenges of getting them on the road.
Stanford University’s autonomous vehicle, “Junior,” was invited to appear at the symposium, but its creators regretfully declined since “Junior” was in the midst of a software upgrade.
|
THE NEED
Highway fatalities number nearly 35,000 annually in the United States, and 1.2 million worldwide. Car crashes cost the U.S. economy $160 billion annually. Time and fuel wasted in traffic congestion add $87 billion. On top of that are greenhouse gasses and other pollution caused by inefficiencies in car travel.
While the idea of driverless cars began as a novelty, it is now seen as a solution to many of these problems. And with computers, lasers, and wireless communication at the heart of these cars, the Bay Area, especially Silicon Valley, is increasingly the place where critical work on self-driving cars is taking place.
This is what inspired Santa Clara Law students Kevin Rogan and Tijana Martinovic, co-planners of the Law Review Symposium, to select autonomous cars as this year’s topic. “The autonomous car, in some form, is inevitable,” Rogan said. “The technology is there, but the law and public opinion need to catch up. And the biggest impediment is the psychology of the driver. We knew these topics would be interesting and stimulate more interest in this area of technology and law” (see “Leading and Learning,”).
While the idea of driverless cars began as a novelty, it is now seen as a solution to many current problems, including crashes, wasted time and fuel, and pollution. And with computers, lasers, and wireless communication at the heart of these cars, the Bay Area, especially Silicon Valley, is increasingly the place where critical work on self-driving cars is taking place.
Many symposium attendees, such as Paul Cott, a robotics engineer, attended to “keep up with industry and law.” Others came to network with a group of specialized speakers and attendees who are leaders in the field. O. Kevin Vincent, chief counsel of the National Highway Traffic Safety Administration (NHTSA), said he expected “great benefits to come from the work of people in this room.”
Dean Donald J. Polden applauded the Law Review for “bringing together such an amazing group of minds in law, engineering, and policy.” He said the event “really demonstrated the Law Review’s foresight in choosing a relatively new topic that cuts across so many areas of the law and at the same time draws on the Law School’s high tech expertise and important location.”
While the idea of driverless cars began as a novelty, it is now seen as a solution to many current problems, including crashes, wasted time and fuel, and pollution. And with computers, lasers, and wireless communication at the heart of these cars, the Bay Area, especially Silicon Valley, is increasingly the place where critical work on self-driving cars is taking place.
THE TECHNOLOGY
Computers began making their way into cars before the 1980s with antilock brakes and traction control systems. Next came navigation systems, parking aids, and cruise control. These automated features are precursors to autonomous vehicles, in which the driver is superfluous.
|
Google has test cars along these lines, modified Priuses and Audis. Stanford does too. But because of California law, all AVs need a driver ready to take the wheel when prompted.
The strong point of the “self-contained” (but reliant on GPS) AV is that it can be put on the road without major changes to the highway system and can share the road with you.
At the other end of the spectrum, says Santa Clara Law Professor Dorothy J. Glancy, an expert on transportation and faculty advisor for the symposium, are “connected AVs, vehicles that are constantly sending and receiving information to and from a wireless network.” They send messages to each other— such as checking in prior to a lane change—and send and receive messages from roadside infrastructure about upcoming potholes, accidents, and weather conditions.
With cars sharing information and coordinating movements, a platoon of electronically linked cars could be traveling at a high but constant speed, nearly bumper-tobumper on heavily traveled commuter corridors, before fanning out to their various destinations.
This “connected” model would require physical modifications to highways, such as dedicated, protected lanes and instant access to a super-reliable wireless system. And how will these polite, communicative cars interact with conventional cars, with their soon-to-be archaic blinking signals and jarring horns?
Symposium speaker Rich Seifert ’06, an engineer and president of Networks & Communications Consulting, is concerned whether the country’s current wireless system can handle what in essence will be “thousands of laptops moving in different directions at fast speeds.”
Seifert warned, “Any vehicular control system that depends on communication systems and won’t work in the absence of one will be a big problem.” Santa Clara Law Professor David D. Friedman concurred. “Wireless needs to be the frosting on the cake,” he said, meaning that all connected cars need to operate also as self-contained cars.
Communications law expert and symposium presenter Mark Johnson, senior associate at Squire Sanders in Washington, D.C., is more confident than Seifert in the capability of wireless technology. He believes the currently available dedicated short-range communication spectrum, developed for vehicle communication and safety, is up to the task of guiding throngs of AVs.
While communications experts, policy makers, and urban planners are assessing the big picture, software engineers grapple with how to put driving into code. Following traffic laws to a T would create a car that might be the slowest auto on the road and sit at four-way stops forever.
Should the driver be able to override the car’s impeccable manners? The Google cars have settings from cautious to aggressive. Will an assertive setting be used against the driver in the case of a crash? Will third-party software be available to create programs to boost speed and performance?
LIABILITY: PUNISH THE TECHNOLOGY?
Not surprisingly, technology is racing forward, but liability issues, regulation, and public opinion are slow to catch up.
Tort and criminal law, safety regulations, and the insurance industry all developed around the concept of a vehicle controlled by a driver. Having the driver in the backseat taking a nap, or no driver, changes everything.
Liability law is likely to discourage automakers from developing AVs, said symposium speaker Gary Marchant, Lincoln Professor of Emerging Technologies, Law and Ethics at the Sandra Day O’Connor College of Law at Arizona State University.
If you take away the driver (who, in part, causes 90 percent of accidents), plaintiffs will sue the manufacturers, claiming design defects (in the words of one attorney general, “Every fender bender will be a products liability case”). The question, Marchant said, will become, “Could this design be done more safely?” This analysis, he says, is very tough on defendant car manufacturers.
New technology, he said, can hurt defendant automakers. “Jurors are skeptical about novel technology and may have visceral reactions against it. We see this in the area of biotech food. Jurors ask, ‘Why are you gambling with people’s lives?’” If harm results from a product that is marketed as ensuring safety, he explained, jurors will “punish the technology.” “There are lots of products that make things safer, such as vaccines and air bags, but the manufacturers still get sued,” Marchant said.
In addition, by the time cases get to trial, jurors have “hindsight bias.” “By that time,” Marchant added, “a better product may be on the road.”
Marchant believes increased liability of manufacturers is a “huge barrier to technology unless we find policy ways around it,” such as limiting liability, which has been done for manufacturers of vaccines.
Liability issues become even slipperier if the vehicle can toggle between autonomous and manual operation. Does liability shift when the driver takes the wheel? As lawyers tossed out worst-case scenarios, Brad Templeton, author of the blog Robocar, asked, “Will the U.S. be stupid, or will liability make it stupid? Singapore [and] China don’t have these legal problems.”
SCU Law Professor Kyle Graham is not so sure that plaintiffs will fare much better in early AV lawsuits. “Early adopters of technology get less favorable treatment” in the law, he said, because they are seen as risk takers. In the early years of aviation, he continued, “if you, as a passenger, took off in bad weather, you were contributorily negligent. With new technology, there’s a tendency to blame the user.
“It may take decades for a coherent and predictable set of rules to develop,” he said. “In the meantime, uncertainty in the law will deter both innovation and lawsuits.”
INSURANCE: RISK AND THE RULES OF THE ROAD
Santa Clara Law Professor Robert Peterson, director of Santa Clara Law’s Center for Insurance Law and Regulation, said he sees “a different picture.”
“The driver will be the first line of liability if the car violates the rules of the road. Owners will have a nondelegable duty, just as they do in California to maintain their brakes. But if there’s a design issue, the manufacturer will be liable.” So, since driver error will be reduced with AV, drivers’ insurance rates should go “way down,” but the cost to manufacturers will go up. The buyer will pay less for insurance but more for the car.
California insurers let out a collective groan when they hear the words “rate change.” Proposition 103, passed by California voters in 1988, has made the process for rate changes, up or down, Sisyphean. (See Peterson’s essay, “New Technology–Old Law: Autonomous Vehicles and California’s Insurance Framework,” on page 12). Peterson suggested that car manufacturers could become the insurers of the vehicles they produce, thereby circumventing Prop 103, which does not apply to commercial insurance policies. “They could build insurance into the cost of the car, as they do with ski bindings,” he said. Kenneth A. Manaster, SCU Law Presidential Professor of Ethics and the Common Good, offered that the legislature could cap liability at a certain amount.
STREET CRED
Nevada has become the first state to legalize AVs, and under Nevada law, AVs aren’t required to have a driver in control (just capable of taking the wheel). So what’s to stop you from sending one to drop off your dry cleaning or take your kids to school? And what’s to stop terrorists from stuffing an autonomous van full of explosives and programming it to their target?
“Autonomous vehicles will be irresistible to hackers and geeks trying to get the cruise control lane to go haywire,” Rich Seifert said. Christopher Kitts, SCU associate professor of engineering and director of the Robotics Systems Laboratory, said, “We can’t assume benevolent use of technology but must prepare for the nefarious.” People may spoof your GPS to give you wrong directions, he said, and hackers may even send erroneous messages that another car is about to hit you head-on at 100 miles per hour.
Can you avoid a DUI by napping in your ride, or do you have to be alert and capable of taking over the controls if there is a system failure? Who will be responsible if a car violates the rules of the road?
Speaker Frank Douma, associate director of the State and Local Policy Program at Hubert Humphrey School of Public Affairs, University of Minnesota, asked whether elimination of the driver excluded all of the “intent” crimes relating to motor vehicle operation. Santa Clara Law Professor Kandis Scott asked whether someone could face criminal liability for overriding safety features.
Can you avoid a DUI by napping in your ride, or do you have to be alert and capable of taking over the controls if there is a system failure? Will those who ride in AVs need driver’s licenses in case they need to take control of a car traveling nearly bumper-tobumper at 60 miles per hour? (Arguably, Kitts said, this requires more skills than currently required for a driver’s license.) Who will be responsible if a car violates the rules of the road?
Santa Clara Law student and engineer Timothy Takahashi asked what would happen if software from other jurisdictions caused cars, or communication systems, to malfunction. Counsel for State Farm mentioned that a Google test car, programmed for California, made an illegal left turn in Nevada.
Kidnappings by hacking into a car’s program is a concern, as are scrambled signals, computer viruses, and having your AV switch to manual when it’s chauffeuring your kindergartner to school.
IF WE BUILD IT, WILL THEY COME?
Autonomous cars have a lot going for them: they offer increased safety, easier commutes, and more transportation options, especially for those who are currently unable to drive due to age or disability.
But transportation specialists have learned that safety does not always trump autonomy and control: the public rejected technology that would require seat belts to be fastened before you could turn on your car. And the public, Glancy said, may be skeptical about GPS-controlled, camera-equipped cars, fearing that they could turn into surveillance devices.
Speaker M. Ryan Calo of Stanford’s Center for Internet and Society, said, “People fear losing control, and they have a different reaction to safety risks outside their control versus inside.” They are willing to take on the risk of driving because they are used to it. And they assume that most drivers don’t want a head-on collision. Can the same be said for most computers?
These fears argue in favor of taking an incremental approach to AVs. Kitts said, “The path toward automation requires giving up some control, and drivers have already begun to do this.” Some new cars already offer lane-change warnings, curve-speed warnings, and “forward collision avoidance and mitigation.”
AVs may also hold promise for increased use of public transportation. People who have already surrendered control to drivers of buses and even driverless airport shuttle trains may be more receptive to AVs. But former urban planner and symposium speaker Bryant Walker Smith, who is a Stanford Law School Fellow at CARS, cautioned that making AVs safer and more desirable will not relieve congestion, alluding to the “if we build it they will come” concept that has proven true with respect to California freeways.
ROAD TESTING
The U.S. transportation bureaucracy can make or break the future of AVs.
Vincent, NHTSA’s chief counsel, reported that the federal government is very interested in “making cars smarter to compensate for human mistakes,” in hope of “eliminating a huge number of traffic deaths, lost time, and economic loss.” But Vincent acknowledged that regulations “need to catch up with technology.”
Before the traffic administration can impose regulations, it is required to engage in a rigorous cost-benefit analysis. “This is our biggest regulatory burden,” Vincent said. “We have to come up with the data.”
Stephen Wood, assistant chief counsel at NHTSA, added that in the current economy there is “a lot of push for costbenefit analysis.”
Just as in the area of insurance, in the world of safety testing, AVs are odd ducks. First off, SCU Law Professor Kerry MacIntosh said, “The technology exists in concept but not in metal and steel that we can test.” There are no manufactured AVs, only conventional cars souped up. Panelist Jon Anderson said, “To find out about safety, sooner or later you have to test the system all together, the connected cars moving together in a fleet controlled via a wireless communication system.” With the technology changing so fast, “How do you regulate a moving target?” MacIntosh asks.
Symposium attendee Dr. Steven Shladover, research engineer with U.C. Berkeley’s California PATH Program (Partners for Advanced Transportation tecHnology), questioned how the safety of autonomous vehicles can be proven relative to conventional cars. “Demonstrating safety will be a daunting task,” he said. “With conventional cars, an average of two million vehicle hours are driven between fatal crashes, and 50,000 between injury crashes. How can AVs be driven enough to provide statistically valid proof of higher safety?” And is it even technically feasible to do that well?
Autonomous vehicles challenge vehicle codes, criminal law, and even international law. Under the Geneva convention, Bryant Smith said, “Every vehicle must have a driver, and every driver must be in control of the vehicle.”
Several commentators expressed frustration at the slow pace of the regulatory process, which is tied to the seven years that it normally takes for auto manufacturers to introduce new cars, and expressed the concern that it is too slow for technology.
|
WHO'S DRIVING, LAW OR TECHNOLOGY?
Beiker, of Stanford’s CARS project, said the symposium was “a great assembly of experts in the field and a fantastic opportunity to learn from them,” and commented, “I am leaving here smarter than when I got here.” And Smith, of Stanford, called the symposium “groundbreaking; the first one this intensely legal.” It made him ask, “Will technology drive implementation or law drive technology?” SCU engineering student Mark Rasay learned that “Concerns of lawyers don’t always coincide with concerns of engineers.”
Cott, the robotics engineer, said the symposium was “very worthwhile. Being a businessman and engineer and not a lawyer, I got a lot of insights.” He said he comes away with the impression that implementation of the AV on public roads will be “in incremental steps, versus big steps.”
In spite of plenty of talk of obstacles to the vision of a platoon of synchronized cars smoothly gliding down freeways, many symposium participants had a sparkle in their eyes when talking of summoning their pod with their cell phone, programming it for errands, and hopping in the backseat for a relaxing commute home.
Susan Vogel is a frequent contributor to Santa Clara Law.
