Tesla Knows How You Drive—Should You Care?

 

In an article in the October issue of the engineering professional journal IEEE Spectrum, Mark Harris investigates the depth and volume of customer-generated data that Tesla acquires every day from millions of its cars on the road.  The reasons for all this data collection appear to be benign for the time being, but it's truly a new thing in the automotive industry, and potential misuse of the data is something to worry about.

 

In common with all other new cars, Teslas have what are called "event data recorders" (EDRs).   Similar in function to an airliner's black box, the data recorder keeps a constantly updated 5-second record of speed, accelerator and brake conditions, steering, and other data relevant to diagnosing a crash.  In the event of a wreck, the last data set is preserved so that investigators can reconstruct the conditions leading up to the accident.

 

But Tesla cars go way beyond the EDR minimum.  Every minute, the car's GPS location and certain other data are recorded.  And when (not if) the car next gets in touch with its designated wireless hub, it uploads an anonymized version of the data to Tesla HQ through the Internet.  Technically, the car's owner is not linked to the randomized ID number that accompanies the upload, according to an engineer under the alias of Green, who has examined scrapped Teslas (as well as the one he owns) to determine what the famously close-mouthed company is doing.  But as Green points out, if you have anonymized data showing that the car leaves a certain residential address at 8 every morning and returns there at 5 every evening, it's not going to be hard to figure out whose car it is. 

 

Besides the location data, the vehicle's Autopilot system can do something called Shadow Mode, according to former AI head of Tesla Andrej Karpathy.  While the human driver is in control, Autopilot pretends to drive the car and compares its steering and control outputs with what the human actually does.  When there's a discrepancy, Autopilot can take a data sample, including camera images and other details, and upload it to Tesla HQ to enable continuous improvement of the Autopilot algorithms.  Multiply this by the several million Teslas on the road, and you have the world's best test bed for improving autonomous-driving software.  This is yet another example of the tech world's powerful largest-network advantage.  Once a player in a networked system gets to be the biggest, that organization has a huge advantage over the other players because of the synergistic effects of network nodes supporting each other, roughly speaking. 

 

Of course, Musk and his engineers say that is the only reason they're collecting all this data:  to improve the Autopilot system.  But it's come in handy in court at least once, when the father of a teenager who died in the fiery crash of his Tesla sued the company.  Tesla was able to present the judge with a detailed catalog of many times when the driver tore around town at up to 130 MPH, establishing that the teen was not driving responsibly. 

 

In fairness to Tesla, they are only doing what any sensible company would do in the same situation.  If Ford or Volkswagen had happened to climb to the top of the U. S. electric-vehicle heap first with an autonomous car, they would probably be doing more or less the same data-gathering.  In principle, even Tesla owners can decline to have any Internet connection made to the car, but no one knows of any owner who has actually done this.  This is probably because the intersection of (people who buy Teslas) and (people who don't want their hardware connected to the Internet) is the empty set. 

 

Should we worry about Tesla, or any other car company for that matter, collecting huge piles of data on where we drive every minute, and how fast we drive, and how safely we drive?  There are two entities that have strong reasons to access this data, and the main concerns may come from them.

 

The first entity is government—Federal, state, and local.  Already, state governments are beginning to wonder how they will keep collecting highway-tax revenue as more drivers turn to electric vehicles, which completely evade the X-cents-per-gallon gasoline tax that has up to now been a mainstay of highway funding.  It's always seemed to me that if you take a libertarian point of view, the people who use the roads should pay for them.  Up to now, it was impractical to tell who was using which road, but as more cars get equipped with follow-me-everywhere software, the technology to assess road taxes by miles used wouldn't be that hard to do.  But for various political reasons, the states seem instead to be leaning toward a flat annual tax on electric vehicles that will more than make up for the lost gasoline-tax revenues.

 

The other entity that would like to get their hands on the data is the auto-insurance industry.  It's not hard to imagine developing algorithms that would take in a year's worth of digital driving data on you and assess a personalized insurance cost that would precisely reflect your driving habits.  This would be very popular for safe drivers and highly unpopular for the other kind.  Of course, as Autopilot and its ilk get better, the insurance companies are going to have to deal with increasing numbers of vehicles driving themselves, and the liability implications of that situation are far from being sorted out.  But it's likely that the insurance industry will develop some kind of certification process that you'll have to deal with in order to obtain insurance on a car with a given type of autonomous driving capability. 

 

Finally, there is the general creepiness factor that some software somewhere knows where you've been.  But as we've gradually gotten used to that with mobile phones, I suppose it won't be much different if our cars know what our phones know already.

 

For now, just being aware that this data gathering is going on may be the most we can do about it.  But while improving autonomous-vehicle software is a laudable goal, it won't be surprising if hackers or other malevolent actors eventually exploit the data stream that Tesla extracts every day from their cars.

 

Sources:  "The Radical Scope of Tesla's Data Hoard," by Mark Harris appeared on pp. 40-45 of the Oct. 2022 print edition of IEEE Spectrum.

When Driving Becomes Illegal

I've been reading a new book about autonomous driving by a couple of academics and the former CEO of Audi in Germany.  They back up their claim of their subtitle—"How the Driverless Revolution Will Change the World"—with detailed statistics, economic analyses, and examples from research labs around the world.  Admittedly, the bloom has somewhat faded from the autonomous-driving rose in the last year or so, but this may be just a temporary lull in what looks to be a decades-long progression of the technology toward full Level 5 totally hands-off robotic driving. 

No one has yet routinely fielded a car that can handle all the roadway and environmental conditions a human driver can deal with, which is the Society of Automotive Engineers' (SAE) definition of Level 5 performance.  But some companies, notably the Google-fostered Waymo, are coming close.  For example, Waymo claims that in 2017, its autonomous test cars racked up over 350,000 miles with an average distance between "disengagements" (occasions when the human driver takes over) of almost 5,600 miles.  Of course, one suspects that few if any of these miles were driven on the New Jersey Turnpike at night during a January blizzard, which is the worst driving conditions your scribe has personally experienced.  Nonetheless, it looks like cars that drive themselves are headed our way.

I remember talking about autonomous cars with a friend one day a few years ago, and I asked him under what circumstances he would give up driving his own car manually.  He answered, "When they pry the steering wheel from my cold dead hands."  I suspect that here in Texas especially, this is not an uncommon sentiment.  The advent of the automobile in the early years of the twentieth century fundamentally reshaped American life, especially for women.  The freedom to pick up and go wherever you wanted, whenever you wanted, without bothering about train schedules or other third-party issues, was fundamental to the growth of the nation and molded much of its infrastructure to this day. 

At first glance, there is no obvious reason why the advent of autonomous vehicles would limit this freedom.  After all, if it amounts to nothing more than a kind of robotic chauffeur, your autonomous car's controls will do your bidding every bit as well as you could if you were driving—probably better, in fact. 

But buried on page 81 of Autonomous Driving is a hint that the transition from where we are now—with manual driving being the norm and autonomous driving the still-rare exception—to an all-autonomous fleet nationally may not be as smooth as you might think.  The reason has to do not so much with the fact that human drivers are bad drivers—most of them aren't.  But a manually driven car will lack the communications and infrastructure connections that are shaping up to be an essential part of the autonomous-car package. 

You may have seen simulations of how heavy traffic in both directions at a four-way intersection can flow smoothly without any traffic lights or stop signs, as long as the vehicles coming to the crossing are able to interact with each other and mutually work out subtle speed changes that allow collision-free traffic flow.  The simulators don't show this, but all it would take to snarl up such an ideal situation would be one comparative-lunkhead manual driver who was unaware of all the intense coordination going on around him, and tried to bull through the intersection like everyone else seemed to be doing. 

Enlarge that snarlup to nationwide proportions, and you can see why the developers of autonomous cars are concerned that the continued presence of old-fashioned manual drivers on the road may some day present a significant obstacle to the progress of autonomous vehicles. 

The authors propose some interim solutions, such as reserving special lanes on some roads for autonomous cars and even prohibiting manually-driven cars altogether in certain areas.  They say that the best solution, however, would be if "the transition to driverless cars is accelerated by means of legislation."  In other words, at some time in the perhaps not so distant future, they hope that legislators will summon the nerve to ban manually-driven cars altogether.  So much for my friend with his cold dead hands.

This plan is not without its aspects of class and socioeconomic implications.  Who drives the oldest, most decrepit cars nowadays, and who is least likely to be able to afford even a used autonomous car if the time comes that manually-driven cars are banned?  Poor people, that's who.  Of course, this notion assumes that the dominant model of privately-owned and privately-driven cars persists long enough to last to the time when Level-5 autonomous cars are widely available and affordable by most people.  There are autonomous-car proponents who claim that car ownership will wither away as cheap robotic Uber or Lyft-type cabs will make public transportation so appealing and inexpensive that even the poorest person will not miss their old junk manually-driven vehicle as they take advantage of the newer, cleaner, and more convenient and reliable autonomous cab services that will show up. 

That may make sense for places like New York or Los Angeles.  But what about small towns and rural areas?  Whatever the costs of autonomous cabs, they will be more expensive if the trips are farther, and for many rural and semi-rural areas, owning your own manually-driven car will probably be cheaper overall than paying for some autonomous-driver cab service that may never show up where you live at all.  Historically and on average, mass public transportation has never been profitable, and the economic case has yet to be made (to me, anyway) that the advent of autonomous vehicles is going to change that. 

Besides the economic argument, there is the principle that when I drive the car that I own, it becomes an extension of my personality, whether for good or ill.  This metaphysical connection between the human being and the vehicle is behind much of the present automotive advertising you see.  It's behind my friend's vehement opposition to letting go of the steering wheel to allow a robot to drive.  And I suspect that if the question of banning manually driven cars ever came up in the Texas state legislature, for example, what would happen next would leave our calm, logical German book authors with their heads spinning.

So the day may yet come when it will be not only inadvisable, but illegal, to get behind the wheel of a car and drive it yourself.  But if it does, it won't be without a struggle.

Sources:  Authors Andreas Herrmann, Walter Brenner, and Rupert Stadler wrote Autonomous Driving:  How the Driverless Revolution Will Change the World (Bingley, UK:  Emerald Publishing, 2018).  An interesting side note which doesn't change my generally positive view of the book:  shortly after it was published, Stadler, who at the time was CEO of Audi, was arrested in connection with the Volkswagen emissions scandal (see my blog on this at https://engineeringethicsblog.blogspot.com/2015/11/vws-in-fix-with-their-fix.html), and released in October 2018 after stepping down from his post with the company.  I also referred to the Wikipedia pages "Autonomous car" and "Rupert Stadler."  A simulation by MIT researchers of an intersection with all autonomous cars can be viewed at https://www.youtube.com/watch?v=kh7X-UKm9kw.

Driverless Cars: Good News for Whom?

After some cautious toe-dipping by Google in a well-publicized series of experiments on public roads in California and Nevada, other more serious players are now eyeing the waters of driverless cars.  According to a recent New York Times report, automakers including General Motors, Volvo, Infiniti, Mercedes-Benz, and Tesla have either already fielded limited-capability "lane-keeping" features in their high-end models, or plan to unveil more advanced systems soon that will allow complete hands-off driving under a wide variety of conditions.  Absent a flood of new restrictive legislation, which hasn't happened so far, it is fairly safe to say that the autonomous vehicle is just a few blocks down the road and heading this way.  Is this a good thing, and if so, for whom?

Danny Crichton, a Ph. D. student in the Harvard John F. Kennedy School of Government, thinks it is.  Writing in a recent issue of National Review, he waxes rhapsodical over the benefits of automation past and future, and has this to say about driverless cars:  "Perhaps no technology has more potential to improve our quality of life than the autonomous car.  We will be able to relax during our commutes, reducing our stress and improving our health.  Autonomous cars could almost instantaneously deliver a greater number of goods and services, such as meals, household supplies, and home-maintenance services, giving us more leisure time."

Crichton clearly writes from a perspective in which driving is just one more daily chore we have to put up with on our way to our real job of teaching or administrating or studying for our Ph. D. from Harvard.  Rather oddly for a person who researches labor economics, he never once mentions an occupation by which about one out of every forty employed persons (2.4%) in the U. S make their living:  professional truckdriving.

If you are a sober, responsible family man (or woman) who couldn't cut the grade in college but want to make a decent living by working hard, truckdriving is one of the more attractive jobs.  Especially with the recent oil boom fueled by fracking technology, truckdrivers have been in great demand.  For a while there were billboards on I-35 in Central Texas advertising large signing bonuses for truckdrivers willing to go to work in the oil fields.  While the hours are long, the work stressful and sometimes dangerous, and time at home is limited, millions of truckdrivers earn enough to support a family.  Many of them are members of minority grouns, and quite a few own their trucks, making them entrepreneurs.  Almost every dump truck I see servicing a construction site around Central Texas has a sign on it with the Hispanic surname of the owner-operator. 

I don't know when, or if, trucking companies will go to autonomous driving systems.  Because of their specialized skills and responsibilities, long-distance and heavy-equipment truckdrivers may be the last cadre of humans to yield the driver's seat to a robot, long after all passenger cars have turned into mobile Internet lounges.  But operators of delivery fleets would like nothing better than to turn their personnel headaches into autonomous-vehicle maintenance accounts.  There remains the question of who or what picks up the package from the back of the UPS truck and carries it to your door, but quadcopters are waiting in the wings for that.  I'm not sure how a quadcopter will ring a doorbell, but by then maybe we'll have wireless doorbells.  Local delivery service is one of the applications that Mr. Crichton explicitly envisions as being done by autonomous vehicles.

Human beings have an obscure but persistent longing for permanence.  If we find a good thing, we want it to go on indefinitely, and that goes for jobs as well as other things.  But it's generally a bad thing to use legislation or union muscle to artificially preserve specific categories of jobs in the face of technological changes.  This kind of thing carried to an extreme produces the antique-car museum that is Cuba, and stifles the increasing technology-fueled productivity that Crichton praises in his article.  If increased productivity means we can do more with less, the economy as a whole benefits, but some people stand more of a chance to benefit than others.

Today's truckdriver in an earlier time might have been my grandfather's iceman, who routinely lugged 300-pound blocks of ice around in a horse-drawn wagon and hauled chunks of it into kitchen iceboxes.  The electric refrigerator eliminated those jobs by the 1950s, but at the same time the trucking industry grew and eventually supplanted rail as a dominant form of goods transport.  And it takes a lot more truckers than it does railroad workers to deliver the same amount of stuff. 

So far, autonomous-driving technology is expanding into what the New York Times terms a regulatory vacuum.  A few states have passed laws either licensing or restricting such cars, but in most states it is still neither prohibited nor explicitly allowed.

Eventually, a driverless car will be involved in a fatal accident.  We may or may not hear about it, depending on the skill of the automaker's PR people.  But whenever such an accident becomes public knowledge, the National Highway Traffic Safety Administration will then receive its legal warrant to examine the whole issue of autonomous vehicles.  The outcome of its study will be critical to the question of whether the technology will continue to be deployed smoothly and cautiously, or whether labor groups such as truckdrivers who feel threatened by it will seize on the incident to mount a crippling regulatory attack that will stop the technology in its tracks.  If that happens, the nation as a whole may end up the loser.

The growth of a new technology is a fascinating thing, bound up in both technical and social issues that can hinge on small but critical events.  The next few years will show whether driverless cars make it big and relieve most of us from what is often a burden—and whether they relieve thousands or millions of professional drivers of their jobs.

Sources:  Danny Crichton's article, "Fear Not the Robot," appeared in the May 4, 2015 edition of National Review, pp. 34-35.  The May 2, 2015 online edition of the New York Times carried the article "Hands-Free Cars Take Wheel, and Law Isn't Stopping Them," by Aaron M. Kessler, at http://www.nytimes.com/2015/05/03/business/hands-free-cars-take-wheel-and-law-isnt-stopping-them.html.  The statistic about the number of professional truckdrivers in the U. S. was from the website http://www.alltrucking.com/faq/truck-drivers-in-the-usa/.  And my grandfather really did run an ice plant for a number of years in the 1930s.