On Sunday, August 9, The New York Times ran an editorial, “Protecting Cars from Hackers,” discussing the recent publicized hacking incidents of Fiat Chrysler and Tesla vehicles, with Fiat Chrysler voluntarily recalling 1.4 million vehicles to fix the bug.
As our cars get smarter, we can expect more of these types of incidents. To be sure, there are new risks presented by the rise of smart cars—particularly when automated systems take over driving task responsibilities previously held by drivers—but the Times’ editorial board’s recommendations will not make us safer. In fact, if we listen to them, we will end up with more highway fatalities and injuries.
The Times recommends:
The National Highway Traffic Safety Administration, which regulates auto safety, insists that it is closely monitoring these new technologies, and is running tests on car software. The agency has also encouraged the industry to create an information-sharing center through which companies can exchange information on security threats.
That’s good news. But N.H.T.S.A. should also start writing basic security standards that require automakers to test the software and make sure a car’s wireless system cannot be used to control the engine and brakes. The agency’s regulations on airbags, seatbelts and crash testing have helped save countless lives. New rules for software that operate cars could prove just as important.
There’s a lot wrong here, so let’s unpack a few points. The Times wants NHTSA to start issuing a flurry of rulemakings on automotive cybersecurity and to “make sure a car’s wireless system cannot be used to control the engine and brakes.” My engineer friends will have already winced at the mangled and incoherent terminology deployed by the editorial writers, but what would prohibiting “a car’s wireless system” from “control[ling] the engine and brakes” mean in terms of, say, self-driving taxis that may be on the horizon? Based on any reasonable reading of the Times’ misguided call to action, it would outlaw them. Not only will automated vehicles likely be far safer, automated taxis would allow more people to live car-free lifestyles, something I thought was supported by the progressive Manhattan elites that populate the editorial board.
The Times then goes onto gush about a looming NHTSA vehicle-to-vehicle (V2V) communications hazard warning mandate, claiming “this could reduce accidents by as much as 80 percent by warning drivers when they are getting close to other cars.” First, the 80-percent theoretical crash reduction only applies to unimpaired drivers, and approximately one-third of crash fatalities are caused by impaired drivers. So, the real safety benefits are far lower than claimed by the Times.
But second, and most importantly, the V2V systems envisioned by NHTSA are likely already obsolete, and will almost certainly be obsolete by the time the mandate is finalized and automated crash avoidance intervention systems become available to consumers. As Professor Alain Kornhauser, director of Princeton’s Transportation Program and faculty chair of Princeton Autonomous Vehicle Engineering, has noted with respect to NHTSA’s obsession with V2V hazard warnings,
If less than the entire fleet is equipped then the 80% diminishes as the square of the penetration rate. So, a 10% penetration avoids less than 1% of the accidents. Greater than 30% penetration is needed to get an 8% accident reduction and one doesn't get half of the 80% until the penetration is greater than 70%. In addition, the half-life of a vehicle is about 10 years. Getting the “entire fleet” equipped in a reasonable time scale would require a nationwide retrofit mandate on existing vehicles, not only new vehicles. Unfortunately, the mandated V2V architecture is likely to be obsolete before the entire fleet is equipped.
As CEI noted in critical comments in response to NHTSA’s advance notice of proposed rulemaking on V2V back in October 2014,
But if [V2V and automated] systems would be completely separated under a proposed rule, the best case scenario for a fully automated vehicle under a V2V mandate aimed at generating driver warnings is that the automaker would be required to install completely useless technology—translating to zero benefits and some non-trivial costs, which would certainly fail a basic benefit-cost analysis. After all, what good is an advanced collision audible warning if a driver has no ability to take manual control of the vehicle in response?
In essence, if NHTSA continues on its current path—and the path apparently supported by The New York Times editorial board—we will end up with a mandate for obsolete safety technology that will likely delay the introduction of far superior safety technology. The result of such regulatory failure is crash events that would not have taken place, and thus injuries, deaths, and property damage that would otherwise have been prevented.
Will we face great cybersecurity challenges as cars get smarter? Absolutely. But the Times’ recommendation that NHTSA greatly speed its promulgation of rules to “help win back public confidence” is incredibly dangerous. Just because some editorial writers believe in the inherent value of regulation and believe the public should also fawn over regulatory agencies does not mean safety should take a back seat to public relations.
The good news for those concerned with highway safety is that few are likely to take the Times’ call to action seriously, given how grossly misinformed their editorial writers are about the state of the technology and the regulatory process. But misinformation, particularly from what is regarded as the nation’s paper of record, has a nasty tendency of gaining footholds in Washington.
Below is an unpublished letter to the editor I sent to the Times on the Monday morning following their deeply misinformed editorial:
Re “Protecting Cars From Hackers,” (editorial, Aug. 9):
You write that the National Highway Traffic Safety Administration should “make sure a car’s wireless system cannot be used to control the engine and brakes.” If taken literally, this would mean effectively outlawing autonomous taxis and other promising safety technologies and applications, which have the potential to save far more lives than the vehicle-to-vehicle communications hazard warnings N.H.T.S.A. is currently seeking to mandate.
Contrary to your claim that “[a]ggressively addressing concerns about the security of car software would help win back public confidence,” it has become clear that N.H.T.S.A. lacks the technical knowledge and administrative capacity to accomplish this lofty regulatory goal in the near-term.
Forcing N.H.T.S.A. to move with excessive haste will not only fail to restore public trust in the agency, it would put lives at risk by imposing harsh political controls on new safety technology development that will likely increase the price of the technology and delay its consumer availability.