Driverless cars aren’t meant to happen — at least not yet
May 21, 2018 | Written by John Sarich
It’s not just that a driverless car was responsible — as much as a piece of machinery can be responsible — for the fatal traffic accident in Arizona. It’s not just that there were 34 accidents involving driverless cars reported in California between 2014 and 2017. And it’s not about the fact that driverless cars are essentially uninsurable. It’s not even about having insufficient data or precise enough programming and code writing — though clearly 200-300 million of lines of code isn’t enough.
An overlooked fact is that driverless cars aren’t meant to be right now because they’re electric.
A key assumption about driverless cars is that they’re going to be electric cars — after all, why would a brand new technology be designed to run on gasoline, which is so clearly passé? Never mind that, by not using gas or “fossil fuels,” we’re going to use electricity that’s made from fossil fuels — unless you live in Florida, where your power is nuclear. Never mind that Bentley just announced its plans to go completely electric — when your average driver is a multimillionaire who doesn’t need to work or go anywhere, you can get away with that. This is about the unfeasibility of widespread electrical power from alternate sources. To reliably get to just 80 percent of demand for electricity with only wind and solar power would cost the U.S. $2.5 trillion.
Driverless cars are going to be electric — that’s a given — so what kind of an electric grid would we need to support them? Infrastructure would be a deciding issue, and while the White House has recently floated proposals for infrastructure improvement, no forward movement is likely for some time, purely for reasons of finance.
A presupposition is that driverless cars, and the ability to manage and track them, are going to require a significant amount of GPS and all the other data-driven technology built into them, so they know where they are and where they’re going. They’re “smart” cars — able to make left and right turns, at the right time and the right place. Theoretically at least, they’ll be able to manage all the different traffic inputs they get and maintain the course they need to go… or they may deviate from it, but they’ll know exactly where they are because of GPS. And GPS takes a lot of power, and yes, even more computer programming and code.
There is an obvious incongruity here. Autonomous cars have trouble going down gravel or poorly paved or marked roads. They’re really great in places with public transportation systems, and well-defined interstates or work routes — until they’re not. In Arizona, for example, an exogenous event that a human driver could have easily avoided led to a collision. And while they’ll be able to move us around in certain ways, on certain routes, I don’t think driverless cars are ever going to give us every single road in the country. So they won’t be ubiquitous by any means.
But back to my realistic view: At the moment, it looks like hype is preceding reality. The whole idea of driverless vehicles being electric begs the question of how we could possibly expand the grid to accommodate all this new energy demand. What would happen to all the gas stations we have today, and where would we all go to plug in our cars, driverless or not? And then, how quickly will the obsolescence of gas-fueled cars happen? If electric cars — with or without drivers — really take off, what is going to happen to everyone who has invested in cars that are suddenly worthless, with technology advancing so fast?
Speaking of technology, with the onset of the Internet of Things a given, there’s a dark coda to this theme that has received too little attention, but is worth deep consideration. Driverless semi trucks — meaning 70,000 pounds of truck and cargo coming at you — are in use. It’s a hop, skip and a jump from there to weaponized driverless vehicles, a single rogue car or a fleet of them… whether they’re hacked by a kid “having fun” in his parents’ basement in the next suburb along from you, or a terrorist operating out of somewhere in Eastern Europe. It goes from being an insurer’s problem to a manufacturer’s problem to a societal problem of the highest magnitude.
With a risk like that looming, we’ll all soon be wishing that hackers could just go back to plain old credit card fraud. And how do you ensure against all that risk? That question still hangs.
In a word, the autonomous car is an unsustainable concept not yet ready for prime time.