Imagine this.
A line of cars is stopped at a red light. Everyone’s
chatting or surreptitiously glancing at their phone, or watching a jogger pass
by. The light turns green. Car number one hesitates as the driver struggles to return
from Pokemon world. Oncoming traffic takes advantage and two cars bang a quick
left. Car number two honks and number one wakes up – they both begin to move,
and further back in the line, drivers stir to life and slowly tap the gas as
the car immediately in front of them begins to move. Like a loose-jointed
snake, the conga line finally progresses, but not many make it through before
the next red. The guys in the rear are disgusted.
Now imagine this.
A platoon of Marines, at ease. The drill sergeant calls them
to attention, then issues a command: “Platoon, forward… March!”
All of the Marines, from the front of the ranks to the rear,
step off smartly, simultaneously. Each Marine in the column trusts that his or her
fellow, directly in front, will step off as commanded. The entire platoon moves
as a single unit. It is a beautiful sight to behold.
The behavior of the slovenly, civilian cars and that of the
polished Marines could not be more different. The difference is that the
drivers do not implicitly trust those in front to move immediately when the
green light lights. Lack of trust begets sloppy performance.
Meanwhile, last week, Ford Motor Company announced that they
will be delivering “fully driverless cars,” with no steering wheel or pedals,
within the next five years.
Now reimagine that intersection. It is 2021, and a line of
odd, bulbous cars are all waiting for the light. Each contains one or more
passenger and no driver. Everyone is chatting or overtly glancing at their
phones, or watching a jogger pass by. While waiting, the vehicles communicate
with each other to verify that there are no substandard (human guided) cars
present. The light turns green, and the entire line instantly begins to move, accelerating
smoothly, akin to a platoon of Marines doing close order drill.
Longer term, all of the traffic lights will have disappeared,
replaced with more efficient and fluid roundabouts. Traffic will flow though
cities and across the land in electric silence, with the ease of corpuscles in
our veins.
Why should we want to do such a thing?
Because vehicles are very deadly, accounting for 33 thousand
deaths and over two million injuries each year. Stop and reread those statistics.
The odds of being injured in a motor
vehicle accident are enormous, many orders of magnitude greater than that of
winning the lottery. While this carnage has been gradually reduced by safer
vehicles, increased seatbelt use, and improved roadways, the human and economic
cost is still enormous.
And while an occasional blown tire or faulty ball joint or
failed brake may cause an accident, the great majority are the result of one
thing – operator error. Robots offer great promise in reducing these
driver-induced accidents.
Another good reason – because today, drive time is often
among our least productive. If your time spent in a wifi-connected, driverless
vehicle could be as productive as your home office, multiplied by millions of
journeys, every day, the productivity boost to our economy would be huge.
What might stand in the way of this vision? The technology?
No, advances in sensors and computing and artificial intelligence are already
capable of basic driverless vehicles. Indeed, Google and Tesla have both
fielded vehicles capable of high degrees of autonomous operation. Five years
from now, those capabilities will have doubled, or tripled, or more.
No, the major impediment to autonomous vehicles will be
regulatory and legal. Will autocar developers have a level playing field in all
fifty states, or will they have to deal with a patchwork of local regulations?
When the inevitable accidents occur, who will be at fault? Will you sue Ford Motor
Company, or the laser sensor vendor, or the artificial-intelligence software developer
who built and enabled your car? The specifics of autonomous vehicle law and insurance
protocols will need to be developed.
There is one regulatory tic which must be avoided at all cost
– the demand for 100% perfection. We must be prepared to accept the occasional autocar
accident, without flinching, in the face of fierce media barrages. Say that
autocars are only 90% effective, but reduce the human carnage to 10,000 deaths,
and one million injured. While the carnage is still significant, the human and
economic costs will have been reduced enormously. We cannot allow the perfect to become the
enemy of the good.
In the end, shall we ban human operators? After all, we are
proven imperfect and dangerous.
I think not. The romance of the open road still beckons us.
The beauty of a downshift and acceleration through an ascending curve should
not be lost.
Robots are OK when we just want to go somewhere. But we need
to be able to express ourselves, to control our trajectory on the road and in
life.
Don’t let that spark be stamped out.