Charles C. W. Cooke writes here about the privacy problems with self-driving cars. As often with privacy issues, the devil is in the details. Cooke begins with a fundamentally valid criticism: most people will not own self-driving cars, but will buy access to them on an ad-hoc basis from companies like Uber and Lyft. The implications for privacy from such a cultural shift are serious and should not be minimized. However, Cooke seemingly goes on to argue that it doesn't matter who owns the hardware, because self-driving is inherently not private:
Regardless, everyone will suffer from the catastrophic loss of privacy. Any network of self-driving cars would, by definition, necessitate total and unceasing tracking of their occupants. I may know how to get to the local liquor store without a map, but my car most certainly does not. To make it there in a driverless model, I'd first have to tell it where I was going, and then it would have to ask the Internet, and the satellites, and, probably, my credit card. To the existing framework we would thus be adding a planet-wrapping exoskeleton with a perfect digital memory. The car, far from serving as a liberator, would become a telescreen on wheels --- an FBI-approved bug, to be slipped beneath the chassis in plain sight of the surveilled. At a stroke, my autonomy would be gone. Without permission from the Web, I would be lost in space. A mere server glitch could render me immobile. The government, should it so choose, could stop me dead in my tracks. Yet again, I would be handing over my self-reliance to the government and to the corporations, and asking, plaintively, "Please sir, may I move?"
This ignores two significant issues. First, traditional GPS (as opposed to the AGPS technology used by smartphones) involves only the purely passive reception of the signal broadcasted by the GPS satellite. Using a GPS device implies "asking the satellite" for information in only the loosest sense; there's no two-way "conversation" and the satellite receives no location data that can be tracked. Second, there are mapping and route guidance applications that use only offline, on-device data, with no need to query a remote server. The standalone route guidance devices that had their heyday in the early-to-mid 2000's, before the wide adoption of smartphones, used these technologies exclusively; they're still available at a $50-$100 price point. They are, in principle, fully adequate to guide a self-driving vehicle.
Self-driving cars will raise many questions that are familiar to FOSS and user privacy advocates. Who will truly own the hardware? Who will be allowed to audit the software? How can we protect consumers from oligopolistic tech companies with continually accelerating access to the intimate details of their lives?
But a fair examination of the underlying technologies reveals that self-driving per se is largely orthogonal to the real threats to our privacy and autonomy, both in and out of our cars. Cell phone location tracking represents an unprecedented intrusion of corporations into our everyday lives; cell phones have also become so indispensable to modern life that even Bruce Schneier, in Data and Goliath, admits to carrying one with him "pretty much everywhere [he] goes." Tollbooth transponders like the E-ZPass have been abused by government agencies to surveil drivers. Modern cars typically contain about a hundred embedded computers that mediate between the human driver and mechanical systems like the engine, steering, and brakes; these systems can be co-opted to the service of law enforcement, lenders, or cybercriminals. Ubiquitous networked cameras, combined with highly accurate facial recognition systems, are increasingly permeating the public sphere in a way that individual consumers are powerless to resist. Ensuring that technological change doesn't undermine cherished freedoms --- regardless of whether the threat comes from the public or private sector --- should be a cause that transcends partisanship. But this will require that we pay close attention to the underlying technological issues --- and entangling those issues with divisive rhetoric about self-reliance and American exceptionalism is at best unhelpful.