,
15 tweets,
2 min read
Read on Twitter
An autonomous vehicle thought exercise... let's start with something really familiar. Imagine a cash strapped city with crumbling infrastructure.
Google (or whoever) approaches the city with an offer: your traffic will flow better if your traffic signals can talk to cars. We know you can't afford to do that, but we can. We're happy to modernize your entire signal system, and operate it for you.
It's a public-private partnership! It's a win-win! The city gets better infrastructure at no cost. Google sells more cars. The city enthusiastically signs on. The corporate lawyers and the municipal lawyers negotiate, and the deal (hundreds of pages of paperwork) is done.
And it works! The new signals are installed and traffic is better for everyone. The signals adjust to traffic flow and to special events. Broken signals are debugged online and fixed quickly. It's great.
The algorithm, of course, is under Google's control. Trade secrets. Nobody cares since the traffic is flowing and the signals are working. And the algorithm does a great job counting - even predicting - how many cars are waiting and adjusting the signals.
Google makes this even better with closer car integration. But of course that only works with Google cars. If you have an Uber or Ford autonomous vehicle, you wait longer. Google sells more cars.
And that algorithm can be monetized. Pay extra, the lights turn green for you. Your car knows -- from your phone's calendar -- when you're late. "You're going to be 5 minutes late," it tells you. "Do you want to arrive on time for only $10?"
You don't really know where that rate comes from. It may be based on the the algorithm guessing, from your email, how important that meeting is. It may be a kind of auction, where you're bidding against others.
(The next killer feature: you can pay to slow others down. Might be useful when you're trying to win a client over your competition.)
Even with all this, the new signals are still more reliable than the old ones. And for the most part the public gets used to all this -- in fact, being able to be sure you're on time feels like a great feature.
There are definitely people opposed. It becomes an issue in the election. But the city can't afford to go back. And, besides, the anti-smart city challenger keeps showing up to campaign events late.
(Oh, and if you're a pedestrian, you have to watch an ad on the beg button to get a "walk" signal, unless, of course, you pay for YouTube Premium on your phone.)
None of the steps in this scenario seem unlikely to me -- there are precedents for all of them. Will it happen? Who knows. I'm not even sure autonomous vehicle tech will be good enough for this to happen anytime soon.
But we know companies will find every way to make money -- if governments let them. And if we don't want stuff like this to happen we need to set the rules up front, and be more skeptical than optimistic when we do.
(And think through the similar scenarios when the same company offers to operate the transit system, or to provide partial data to the police and the DA, or…)
