Once upon a time, everybody seemed to believe autonomous vehicles (AVs) would soon be wheeling themselves merrily on every road.
That didn’t happen.
Though we still expect these things to arrive on streets eventually, how they do so won’t be straightforward, and it’s questionable whether private ownership is even a desirable aim. Shared AVs seems like a more promising approach.
Even before we get to that, here are six problems the Apple Car – and anyone else developing such vehicles – must solve.
Saving the world one car at a time
The car you drive is a giant chunk of steel, iron, plastic, aluminium, glass, rubber, precious metals and more. Not all of these materials are as readily available as they once were and the energy used in manufacturing a vehicle is an estimated 2.7MWh/car, according to the European Automobile Manufacturer’s Association (ACEA).
The industry is slowly shifting to make more use of recycled materials, but the inconvenient truth is that replacing the billion vehicles on the world’s roads with EV/Autonomous equivalents would demand more use of recycled materials and almost certainly challenge the supply of those raw commodities. Stellantis CEO Carlos Tavares recently said he doesn’t think there are enough raw materials available to replace all the cars.
What does this mean?
Just as iTunes became a subscription service, it suggests expectation of car ownership will be replaced by acceptance of a variety of models around access, including public transit, bikes, subscription, and ride-sharing schemes.
How much energy do you want to use?
Wired recently claimed Waymo Jaguar generates 1,100 gigabytes of information every hour. Multiple that by 1.4 billion (the number of cars in use worldwide) and that’s roughly 24 trillion gigabytes of data generated daily.
While some of this information is of little value, it must still be processed — with some data kept for training, as proof in case of accident, and for quality control.
Think how much energy and water the data storage centers handling this information will need. Ironically, it could mean that while the environmental impact of driving your car may be lower, the overall impact of manufacturing it and the data storage and servers upon which it relies will cast a nasty, dark shadow across that EV green glow.
Now, consider the energy used by all that street furniture as it, too, shares information and data with the traffic management system. It all adds up to energy which has to come from somewhere — but where?
In comparison, human drivers use around 24 watts of calorific energy for driving, the equivalent of eating a carrot.
How well do you want to be known?
The data lifecyle is another matter. Much of it will be low-grade information that loses relevance fast. But not all of it.
So what data should be kept?
There are limits to what isn't retained. For example, in case of an accident, investigators will need to be able to access all that information to verify the autonomous device kept to the rules of the road, which suggests both vast data archives and a danger that personal information might be identifiable. It has to be identifiable to link back to the vehicle being investigated.
Where data has identity, there is money to be made. This is why I’m certain some data brokers would love to be able to exfiltrate information about how you drive and where you go to add to their existing files of insight into your personal habits.
Think how data brokers already use ‘fingerprints’ (qualities unique to your devices) to track you online. Will they do the same thing for you and your car? Of course they will, unless they are stopped.
Who knows where you are (and where you go)?
But of course, this data is also potentially super-useful. Think how it would feed into improved city infrastructure management.
Advocates of the smart city revolution argue that it should make travel faster and cheaper, optimize public transport, and find the golden fleece. (OK, not the fleece.) One thing it will almost certainly do is enrich a handful of commercial companies developing smart city management solutions.
But what happens to that information if companies go bust or get hacked? How swiftly does the existence of the data turn into intrusive personal surveillance or nasty attacks on individuals by hostile nation states?
That’s even before you consider inevitable lock-in problems when software providers demand stiff exit fees from cities choosing to use other services in order to keep their data.
What happens when software providers quit?
There is also the danger of incompatible data formats making it hard or impossible to move between different systems. We’ve seen this happen before in tech, and in a nascent and weakly regulated new industry it's inevitable we’ll see it again.
Road traffic conditions will improve until the software developer goes bust or a nation state forgets to pay. We will no doubt see how some of these factors play out across smart city deployments in the Middle East, such as at the King Abdullah Financial District Development & Management Company (KAFD DMC) in Saudi Arabia.
Smart cities are big business and are expected to become a $517 billion market opportunity for smart city solutions providers. With so much to gain, competition will be fierce, M&A activity rapid, and the dangers of data lock and incompatibility inevitable. As I said, we’ve seen this before.
How to kill your enemies the easy way (and deny it)
World War I began in 1914 following the assassination of Archduke Franz Ferdinand of Austria. The first attack took place as the royal convoy drove through Sarajevo when an assassin threw a grenade but missed. The second successful attack took place when the vehicle was stalled in traffic on a narrow street. Another assassin shot and killed the Archduke, precipitating a war that claimed 40 million lives.
It is the car that makes that history lesson relevant here. Now, imagine a hostile nation state or well-resourced terrorist band had access to the whereabouts of the Archduke on that day or had managed to hack its way inside the traffic management system to take control of the car’s movements.
They could have chosen where to stop the car.
Harm is so much easier when you control your victim’s environment. That’s not really implausible, given the cost of those insidious, reprehensible, and ongoing NSO Group attacks. Assassination by autonomous vehicle is just a logical next step.
Cars are complicated
Cars are complicated. They're embedded across every part of daily life, and replacing them exposes profound questions that must be answered before any autonomous vehicle project can deliver the degree of transformation we expect.
With its focus on privacy and stance on renewable energy, Apple may have some of the answers to some of these questions. But answering all of them means those Apple Cars will be waiting in the test lab for a while.
I still think Apple should make a bike.
Please follow me on Mastodon, or join me in the AppleHolic’s bar & grill and Apple Discussions groups on MeWe.