In the June issue of Reason, one of my favorite publications, Greg Beato has an article discussing the public policy implications of autonomous vehicles, such as Google’s Self-Driving Car. While I appreciate libertarians (being one myself) taking this technology seriously, Beato makes a number of questionable assumptions and outright factual errors in the piece. Here’s my quick attempt to address some of them.
Beato begins with obligatory Google-bashing common among techies, who seem to either love Google or despise it. (This is probably too simplistic, but this is how it looks like to a Silicon Valley outsider.) The legal issues with respect to Google’s collection of unprotected Wi-Fi data are complicated from a libertarian perspective, those related to Google’s settlement with the FTC “for bypassing privacy settings in Apple’s Safari browser,” as Beato puts it, are not. No one’s privacy was ever violated. All Google was guilty of, as technology policy and privacy analysts here at CEI noted at the time, was
failing to realize a software tweak by Apple rendered one of Google’s help pages inaccurate. There is no evidence that any users were “taken in” or harmed by this inaccurate help page, nor does the FTC allege that Google knew or should’ve known that its help page was wrong. A four-commissioner FTC majority even admitted that Google’s alleged wrongdoing didn’t last very long or earn the company much money.
This is hardly the privacy-invading sin Beato implies it was, but he is obviously setting the stage for his arguments for additional public skepticism of autonomous vehicle technology.
But much of the beginning of the article focuses on the huge potential benefits of vehicle automation, which are large and which we at CEI have highlighted in the past. But at the halfway point, Beato drops this:
But is everyone really so eager to see the automobile, which stands as one of history’s great amplifiers of personal autonomy and liberty, evolve into a giant tracking device controlled by a $250 billion corporation that makes its money through an increasingly intimate and obtrusive knowledge of its customers?
Beato is correct that the automobile is one of the great technological liberators of mankind from the time-consuming drudgery that was previously associated with personal mobility. But the implication that Google is intent on destroying privacy protections by deploying a mobility-enhancing technology is over the top. Autonomous vehicle users in the future, just like users of any digital technology that transmits telemetric data, will be opting in. Google and other potential providers, in turn, will likely be responsive to privacy concerns. The real concern is the ability of law enforcement and other government bodies to access this private information.
But Beato instead makes questionable assumptions regarding technology that is not yet available to consumers, always a dangerous tack to take. For instance, Beato claims, “Even if it were possible to operate the car in some kind of ‘manual’ mode, you would likely still be sending information back to headquarters.” “Even if”? “Likely”? As far as the ability to operate a fully autonomous vehicle manually (i.e., not in autonomous mode), this will be standard. In fact, more than one of the four (not three, as Beato incorrectly states later in the article; they are Nevada, Florida, California, and Washington, D.C.) jurisdictions that recognize the legality of these vehicles (not where they are legalized — more on this in a moment) explicitly requires this feature.
Immediately following, Beato sets up this rather paranoid hypothetical:
In time, Google will know when you arrive at work each morning, how many times a week you go to Taco Bell, how long you spend at the gym. As illuminating as our searches and other online behavior might be, there’s still some room for ambiguity. Maybe you’re doing all those searches on “brain tumor” because a relative is sick, or you’re doing some sort of report, or you’re simply curious. Combine that info with the fact that you start visiting the hospital every week, however, and Google knows you’ve got cancer.
The driverless car, in short, is a data detective’s dream, a device that can discern when you get a new job, how many one-night stands you have, how often you go to the dentist. As demarcation lines between the real world and the virtual world continue to blur, autonomous cars will function not so much as browsers but links, the way we get from one appointment or transaction opportunity to the next. In theory, Google will determine the route to your desired destination based on distance, available infrastructure, and current traffic conditions. But what if Google, which already filters cyberspace for you, begins choosing routes as a way of putting you in proximity to “relevant content”?
More scary “ifs” and questionable assumptions. What if all Google wants from this technology is to create the best road atlas in history? That hypothetical isn’t nearly as frightening from a privacy perspective, but it’s equally as valid as Beato’s series of potential scenarios.
He then makes a reasonable, and I believe absolutely correct, point on consumers’ ability to opt-in: “Many people will no doubt love such new functionality. Others will opt out, or refuse to opt in. Others, however, will simply want to stay as far away from self-driving cars as possible.” It is important to remember that frontier and disruptive technologies, especially those that face public skepticism, generally aren’t mandated overnight. We may in the future wish to create dedicated autonomous-only expressways or develop vehicles that only operate in autonomous mode, but the potential future push to mandate all vehicles be autonomous is a long way off — assuming it will ever happen.
Finally, I need to address what seems to be the most common error made by members of the press and other commentators. Beato claims “[c]urrently it is legal to operate a driverless car only in Nevada, California, and Florida.” This is false in two respects. First, like I said above, four jurisdictions in the U.S. currently recognize the legal status of autonomous vehicles. But to the bigger point, and bigger error: autonomous vehicles are already legal in the United States. In a November 2012 study, “Automated Vehicles are Probably Legal in the United States,” Stanford researcher Bryant Walker Smith analyzed the legal issues surrounding autonomous vehicles and concluded that since they aren’t explicitly addressed, the law does not prohibit their use. The distinction here is between “legalizing” autonomous vehicles and “recognizing the legality” of autonomous vehicles. To be fair to Beato, he is not the first libertarian commentator to make this mistake, but hopefully he will be one of the last.
Autonomous vehicles face many hurdles, both technological and political. While Google may be optimistic in their prediction of consumer-ready models in three to five years, it is certainly a possibility that the technology will be available within a decade. The potential benefits of autonomous vehicles are enormous, as Beato highlights. And, yes, there are plenty of things politicians can do to screw up deployment and adoption by consumers — some of which have already been tried, as I noted last year in The Washington Post. But it is not at all productive to make dire claims about the technology when it is not even available for public inspection, let alone consumer operation. I’ve said it before, but I’ll say it again: Let’s just wait and see, shall we?