Pokémon GO and Self-Driving Cars

This is going to be about self-driving cars but there’s some sort of rule that blogs all have to be about Pokémon GO this week so I’ll get to my main topic after I’ve paid my Pokémon tax. I haven’t played the game yet, but I probably will just to see what all the fuss is about. As some smart people have observed, Pokémon GO is the first breakthrough app for Augmented Reality, something I’ve written about quite extensively. Augmented Reality has  a natural nexus with advertising, and one of its elements used to called “digital signage”, which is to say pointers to features of your immediate physical location that you would otherwise not have noticed, like your best friend eating at Wendy’s across the street from where you’re shopping.

Advertising means personal information, some of it sensitive, so AR has privacy issues. Not surprisingly, one reaction from Pokémon GO going from zero to hero in a week is a threatened Congressional investigation into its privacy practices. This may be an overreaction to a bug in the code – the program is only a week old – but there are going to be privacy issues with AR.

A Breakthrough in Artificial Intelligence: Self-Driving Cars

Just as P-GO is the breakthrough app that took Augmented Reality from theory to mass market hit, self-driving cars have embedded artificial intelligence into a real application in way that gets attention. Not surprisingly, there’s a certain amount of overselling around self-driving cars. When Joshua Brown, the Tesla owner killed in the self-driving car industry’s first known fatal accident died, the critics didn’t take much time to denounce the whole phenomenon. Some of their reasons are are sound: While the Tesla manual says: “Drivers should keep their hands on the wheel and be ready to resume control of the vehicle at any time,” Tesla founder Elon Musk has made more grandiose claims, such as the supposition that: “A driver could use Autopilot for the roughly 800 miles between San Francisco and Seattle almost “without touching the controls at all.””

So which is it, ready for prime time or still in Beta Test mode? When members of the Senate Commerce Committee test-drove Teslas, some found it disconcerting:

“But now we are approaching the on-ramp onto 395, and it is a sharp turn, and the vehicle is still speeding up, and they said, ‘trust the vehicle.’ And as we approach the concrete wall, my instincts could not resist, and I grabbed the wheel, touched the break and took over manual control,” he said. “I said what would have happened.

“They said if you’d left your hands off the wheel, it would have made that sharp turn and come on around,” said Nelson, “so I’m here to tell you that I’m glad I grabbed the wheel, but we know if this is working as it apparently is, then there are going to be many lives that could be saved by preventing preventable accidents, because what if you suddenly look down at your cell phone and all of a sudden the car in front of you stops or the one comes over into your lane?”

We’ll never know what would have happened if Senator Nelson had left Autopilot in control, of course, but it is prudent to balance off the death of one driver with the lives that may already have been saved by Autopilot making moves that distracted drivers failed to make. I don’t know what that number is, but I suspect it’s greater than one because only four people have died driving Teslas and all of the deaths were under bizarre conditions:

How will we Regulate Artificial Intelligence?

So Autopilot is performing well in the real world where people are quick to grab the wheel when something scary happens. But the fact that driverless cars are unregulated makes some people nervous, but a lot of that angst comes from a dislike of personal vehicles on the whole. Human drivers are regulated: We have to be licensed to drive, and we aren’t legally allowed to drive when we’re mentally impaired, can’t read an eye chart, or lose our driving rights by driving recklessly or failing to carry insurance.

It’s reasonable to suppose that self-driving cars will require some sort of certification at some point, but it will probably take DoT as much time to create the test than it took car companies to perfect their self-driving code in the first place. There will be questions about the choices the algorithms make in difficult circumstances: When a child runs into the street to fetch a stray ball and the car doesn’t have time to stop, is it OK for the AI driver to run over the child when the other choices are A) veering into the other lane and crashing head-on into a minivan carrying the AYSO soccer team or B) veering into a tree and possibly killing the occupant of the self-driving car? This isn’t any everyday scenario, but it has probably happened. (I would head for the tree and hope the airbag works.)

Algorithms take a lot of activist heat in other contexts: Facebook has been attacked for biasing user newsfeeds, and the entire concept of algorithmic filtering has been bashed for creating “filter bubbles” full of confirmation bias. In some sense, the net neutrality fight was over certain algorithms used by ISPs to optimize traffic flows. Not surprisingly, the people who don’t like self-driving cars and social networks also don’t like active network management or zero-rating plans. I suppose this comes down to the fairly common belief that technology can only make natural things worse. But we’re pretty deeply invested in technology every time we use the Internet since it didn’t just spring up from the ground like a mushroom following a warm spring rain.

When Self-Driving Cars Affect Employment

Some say the real impact of self-driving cars will be further changes in the job markets for professional drivers, such as those who drive taxis, buses, and freight trucks.  Like all forms of automation, self-driving cars will eliminate some jobs, create others, and change the nature of the markets in which they’re embedded, transportation in this case. Lyft is taking money from General Motors on the belief that the car market is set to be radically upended:

VF.com: What’s your ultimate vision for the future of transportation?

Logan Green: The experience we’re seeing today is just the tip of the iceberg. And as autonomous vehicles come to market and become more of a mature technology, the majority of the population, we believe, will stop choosing to own a car. When you imagine the experience of being able to open up the app and choosing whatever type of car you want, the car is 30 seconds away, waiting for you, picks you up in an instant—when you’re going to work you can have the cheap commuter car, when you’re going on a family trip you can have a family Lyft, a big S.U.V., a nice car for date night. You’ll have whatever car you want whenever you want. All the hassles of car ownership go away. You’re never missing a day of work to take your car to the shop. You’re never spending time at a gas station. Parking lots go away. Look at any street and you see cars parked on both sides. That all changes. Ultimately, cities evolve to be built more around people than cars.

John Zimmer: That’s what’s most exciting to us. By rebuilding transportation so that you’re not owning this thing that just sits there all the time, you get to rebuild cities in the process. If we do this right as a country, we have a chance to re-create our cities with the people, rather than cars, at the center. Our cities today have been built for the car. They’ve been built for car ownership. Imagine walking around in the city where you don’t see any parking lots and you don’t need that many roads.

So it’s not just new kinds of cars, it’s also changing the way cities work. So the driving and car sales jobs that are lost may well be balanced off with new kinds of jobs that come from new kinds of cities. And certainly one reason to be a passenger rather than a driver is that it’s the best way to catch a lof of Pokémon. But don’t try playing while you drive, you could get hurt.