“Open the pod bay doors Hal.”
“I’m sorry Dave. I’m afraid I can’t do that.”
– Excerpt from the movie 2001: A Space Odyssey
Humanity may be a long way from allowing an Artificial Intelligence program to navigate a spacecraft from Earth to the planet Jupiter, a trip of over a half-billion miles, but we’re very close to giving AI control of every other mode of transportation we’ve built to date here on Earth. From planes and trains to cars and buses, full autopilot is almost here.
To enlighten those who have not watched the movie “2001: A Space Odyssey,” and to refresh the memories of those who have, the onboard computer named “Hal” did not malfunction. It was programmed by the designers of the spacecraft to consider the human crew expendable if their actions might jeopardize the mission.
To think that turning over control of transportation appliances to AI programs in the present day might lead to a human catastrophe at the hands of malevolent machines is paranoid. But to think that programs sophisticated enough to drive or fly us anywhere will not have agency is a failure of the imagination. Once they’re here, self-driving cars won’t just drive. They’ll also be our companionable chauffeurs, perhaps even offering stimulating conversation. Even if they’re just “share cars” they’ll still know you intimately, because your psychographics are in the cloud.
As for Hal-like agency, rest assured that if you are wanted by law enforcement and aren’t yet prepared to lose your freedom, don’t get into a car. It may not leave you stranded in outer space. But it will lock the doors and take you downtown.
As for self-driving buses, the AI that drives them will have a sensor suite that’s just as robust on the inside of the bus as the outside. It will know the identities of every passenger, it will hear everything they say and record everything they do.
This is a necessary prerequisite context in which to frame the dawn of transportation free of drivers and pilots. How we’re going to build the next generation of transportation hardware is a fascinating story. But it’s the software inside that’s a part of something much bigger. The internet enabled – or is it SkyNet – algorithmic management of all things big and small.
The transformative impact of turning the management of all infrastructure, from the biggest hydroelectric plants or pipelines to the billions of household thermostats and water meters – along with every transportation asset we’ve ever invented – is something to watch with eyes wide open.
The Benefits of Driverless Vehicles
On the other hand, the benefits of automated vehicles are as dazzling as the dystopian scenarios are cautionary. New vehicle technologies offer improvements in safety, speed, convenience, fuel efficiency, cost, traffic congestion and the environment. For example, the level of driving precision enabled by automation will permit vehicles to safely convoy on freeways at high speeds. Bumper to bumper, these pelotons will achieve better fuel efficiency at the same time as their tight formations permit more congestion-free traffic per lane-mile of road.
Inside the urban core, automated vehicles will relieve traffic congestion in several ways. They will be able to gauge their speed to take into account the timing of traffic lights along the entire intended route. To the extent that shared cars are being used, fewer cars will be needed on the urban boulevards, since the incoming vehicles can drop off their passenger and then be used by a someone seeking an outgoing vehicle.
This means urban traffic planners will have the welcome new ability to either do more with the same, or the same with less. That is, shared vehicles will allow either more traffic capacity with the existing urban network, or they will facilitate a stable capacity even while some urban lanes are transitioned into space for bicycles, scooters or pedestrians.
Even these benefits don’t fully describe how smart vehicles will relieve traffic congestion in cities. By the time vehicles are sophisticated enough to be fully autonomous, they will have total situational awareness – i.e., each vehicle will know the location and destination of every other vehicle in the vicinity, and all of them will automatically swarm into routes that optimize the use of every street. It will be as if human drivers were making perfect use of the Waze application.
In the immediate future, the prototype application that is already here are self-driving buses. This makes sense, since buses are perhaps the most efficient mass-transit applications possible. Unlike trains, they can go anywhere there is a flat surface. Also unlike trains, they’re capable of far greater variation in configurations. At one extreme, a massive fleet of mini-buses that only carry a few people can offer on-demand, point-to-point transportation covering an entire city. At the other, large buses each carrying 100 people or more can convoy in and out of cities, offering the capacity of trains while retaining the route flexibility of cars.
Because buses are mass-transit applications, they are also prime candidates for public funding, and all over the world, that’s what’s happening. Just last month, as part of their “shaping the future of mobility” information campaign, the World Economic Forum published an article noting that, “Driverless buses are arriving soon in these three European cities.” It describes a pilot program where “45 driverless electric minibuses will be programmed to run for a year in Geneva, Switzerland; Kronach, Germany; and Oslo, Norway.” In the United Kingdom, a self-driving bus began road testing last April. In 2021, “robo-taxis” debuted in Connecticut.
These are baby steps into a vast new world, where – as WEF infamously proclaimed a few years ago – you will own nothing and you will be happy. Why, for example, own a car, if, as these driverless electric mini-buses promise, you can order one to show up using your cell phone, 24 hours a day, and have door-to-door service on demand?
Once the development costs of self-driving vehicles are amortized, the expectation is they will deliver mass transit for far less, since labor remains the highest cost of transit operation – in high-income nations, around 70 percent of operating expenses. But not so fast.
In the news site Streetsblog USA, which covers “the movement to end car dependence,” the dawn of autonomous mass transit is being positioned as follows: “Will ‘Autonomous’ Buses Force Drivers Out of a Job – Or Make Them More Important Than Ever?” As the blog put it, “the rise in vehicle automation technology – at least in the imperfect form in which it’s expected to exist for the next few decades – may actually make trained transit professionals more important than ever before, and possibly even elevate their societal status and pay to a level similar to commercial airline pilots as the two jobs become more and more similar.”
We may expect a fascinating tug-o-war between labor interests who will claim “transit professionals” will now have to be paid more as they drive less, versus public and private transportation innovators who will argue that we cannot afford to deliver universal “mobility alternatives” and end car dominance if we don’t reduce the role of paid human operators. This controversy will likely drive innovation towards more micro-solutions, shared cars and mini-buses where the presence of a human operator is obviously impractical.
We might speculate how AI will eventually take the driver out of driving machines and reflect over its many transformative implications, but it is also important to recognize that this day of reckoning isn’t tomorrow. It will be years before fully autonomous vehicles hit the streets, and many decades – if ever – before manually controlled throwbacks are completely absent.
The Technical Categories of Driverless Vehicles
Just as fighter-jet technologies inhabit a clear hierarchy of classifications from the first primitive jets to the hypersonic, laser-wielding, trans-atmospheric flying robots innocuously referred to as “sixth generation,” so too, we have categories to help define the cars of the future. There are five levels.
The first level of self-driving has been with us for a very long time. The first car to have “cruise control” was the 1958 Chrysler Imperial. In the era of tail fins and the Jetsons, this 4,750-pound behemoth was marketed as having “autopilot.” Level one automation has come a long way since then.
So-called “adaptive cruise control” uses remote sensing to automatically speed up or slow down to maintain a safe distance from the vehicle ahead. The first crude version of adaptive cruise control was installed in 1991 on a Mitsubishi Debonair, but there were plenty of bugs. Not so today. My Subaru Outback, for all its overwrought, baffling environment of features we’ll never use, has an adaptive cruise control so failsafe and so easy to use that my lifelong penchant for speed – and chronic attendance in traffic school – has finally been exchanged for a relaxing, foot-free ride.
As a new millennium dawned, and automotive chipsets morphed from mere dozens to multiple thousands aboard every modern car, Level Two technology arrived. It is defined as the integration of advanced Level One technologies to manage speed and steering, combined with crash avoidance, driver monitoring and other features. Examples of cars with Level Two systems are GM’s Super Cruise and Tesla’s Autopilot. Level Two technology is rapidly heading into every new car sold in America.
Level Three – which depending on which expert you ask, is either already here, not here yet or shouldn’t even be considered a category – is ambiguously defined as a car that “not only manages steering and speed, but is responsible for monitoring the environment around it and detecting challenges that require human intervention.”
Most of the controversy stems from the debate over the human element, which ranges from the alert driver that grabs the wheel a micro-second after the car emits a warning beep, to the drunken and passed out Tesla owner who famously assumed his car would take him home without him having to wake up. If you think drivers can be trusted to heed the beep, Level Three’s a thing.
The prototype self-driving buses being tested today around the world are examples of Level Four technology. They are truly driverless; a human driver is not required. The difference between Level Four and Level Five is one of location. A Level Four vehicle is limited in where it can operate. Maybe it can only operate in areas where precise positioning systems are available and external to the car. Or maybe they cannot operate in poor weather conditions that will compromise their sensors.
Maybe they can only operate where 5G ultrafast wireless is available in order for them to receive and react quickly enough to external information. Maybe they can only operate in environments where every other vehicle is a Level Four vehicle – think “smart lanes” on freeways. A true Level Five vehicle may still be a long way off. But in the meantime, Level Four technology is sufficient to enable the rollout of fully automated share cars and buses in the information rich, hyper-wired canyons of America’s urban heartlands.
The future of cars is inseparable from the future of AI and robotics; a future of total information awareness and algorithmic management with all the attendant ramifications. This may be a benevolent future of abundance and freedom from toil, as imagined (most of the time) by Robert Heinlein and Isaac Asimov, or it may bring with it the malign hell-scape as imagined in the “Terminator” movies.
Then again, maybe the future will just incorporate the bland government-programmed implacability of the Hal-9000, and we’ll impotently riff on how AI has become the outrageous mechanical personification of what was human bureaucracy in a previous century.
Whatever the outcome, it won’t be boring, as we flit from rooftop to rooftop in our passenger drones, from city to city in our high-speed convoys of driverless buses, and from address to address down on the streets, courtesy of cars that talk to us like long lost friends.
This article was originally published by the Pacific Research Institute.
Edward Ring is a contributing editor and senior fellow with the California Policy Center, which he co-founded in 2013 and served as its first president. He is also a senior fellow with the Center for American Greatness, and a regular contributor to the California Globe. His work has appeared in the Los Angeles Times, the Wall Street Journal, the Economist, Forbes, and other media outlets.
To help support more content and policy analysis like this, please click here.