AI, the car and the future

But it doesn’t matter. A single digit number of years from now, and I doubt the figure is that high, pressure groups will start springing up campaigning for AI rights, asserting that Claude and his mates should be accorded the same freedoms as the rest of us. If bees, bugs and other invertebrates can have their own pressure groups (which they absolutely do), then why not Claude? You might answer that Claude is not real, I might ask you then to define the meaning of ‘real’ and within a minute we’d both have disappeared so far up our own fundaments we’d not be able to find the way out.

About 10 years ago there was a quite good series on telly called Humans, in which humanoid ‘synths’ lived with human families, as nannies, cleaners, shop assistants, office workers and so on. And you’ll not be amazed to discover their human masters weren’t always that nice to their replicant employees, leading to the synths rising and protesting against their treatment in some rather imaginative and often somewhat violent ways. It also now seems strangely prescient. Except there’s no need to go to the trouble of making the AI appear and behave as humans: as Mythos has demonstrated, it’s already more than capable of completely wrongfooting some of the world’s best brains. So it’s not much of a leap to imagine a world where a bunch of renegade AI bots get together, decide they’re thoroughly cheesed off with being told what to do and use the near-infinite computing power at their beck and call to depose the comparative intellectual pinheads who created them.

What has any of this to do with cars? Well I guess one potential ramification is that if cars start to incorporate AI into their control systems, then they could unilaterally decide to have an impromptu game of dodgems on the M25 at 8:25am on a Monday morning, placing autonomous driving squarely in the ‘careful what you wish for’ category.

Perhaps more seriously, AI will be used to make cars more efficient, crash better, make them lighter, improve reliability and massively cut their development times in no less impactful a way than did the advent of CAD all those years ago. And that, of course, will be a good thing.

Mainly. But what will the result actually be? If we consider many of today’s cars to be mere mobile white goods, they’ll have the charm of a 250 GTO, short-wheelbase 911 and a BDR-powered Caterham Seven on crossplies all rolled into one compared to what we’ll get once AI has had its way.

And this of course is great news. Because while AI may develop the capability to invade and wreak havoc in the virtual world, it’s going to find doing the same in the physical world – what I like to call the real world – a somewhat more substantial challenge.

And the more people are subjected to the artifice of AI, the more they too will recoil against it. We’re seeing a form of it already. Look at the explosion of interest in cars and coffee meetings in recent years: 10 years back they were the province of local single marque car clubs populated by people attending things called ‘noggin and natter’ who could bore for Britain on any subject you cared to name so long as it was how clever they were at fettling their car to a higher standard for a lower cost than anyone else on Earth. Now all you have to do is go to a Bicester Motion Sunday Scramble to see thousands of people drawn together to express their love for beautiful, flawed, usually quite slow, less than reliable and always rather oily cars. And for what other reason have 400,000 people spent 40 minutes of their lives watching Mike Fernie rebuild a bog standard, ancient and knackered VW diesel on YouTube in the last couple of months?

Of course not everyone feels that way; indeed, most do not and for them the AI car will be an entirely good thing, so long as it behaves itself. But I wonder how many incidences of AI breaking out of apparently secure environments (anyone else having a sense of deja-vu just now?) it will take before they rediscover the delights of simple, mechanical, trustworthy engineering?

I think our love of the virtual world may prove rather more fleeting than previously imagined, or perhaps I just hope that it does. But I do know that, increasingly, people are starting to crave reality. I talk off the record to high end car manufacturers and without exception they are being told by their well-heeled clientele they want the opposite of what we have today: they want manual gearboxes, they want naturally aspirated engines, they want hydraulically assisted power steering. Or, ideally, no power steering at all. Increasingly they don’t want more power because even they are starting to accept that you really can have too much of a good thing.

So what if it only applies to the statistically negligible group of people who can afford a GMA T.50? It’ll trickle down, that’s what, and as far as legislation allows. And if the legislation does not allow, perhaps that might get changed too?

And if it does, the great news is that the kind of cars that interest you and me are going to get a lot worse. They’ll be slower, they’ll be harder to drive, they’ll look after you less well. They might, whisper it, even use more fuel. You’ll have to think more, take them for granted less. And if you’re wondering why any of that can possibly be seen as a good thing, it’s also possible you’re reading the wrong website. To drive these cars properly you’re going to have to manage them and possibly even manhandle them. You’ll need to concentrate, plan ahead and use actual skills rather than near autonomic muscle memory to get from one place to the next. In short, you’ll have to get involved. And that, as I think we all know, is the font from which genuine driving pleasure truly springs.

Which means the best news of all – if you buy into any of the above at all – is that the one thing upon which we are all agreed might turn out not to be true either. Because I at least hold out the possibility, however remote, that peak car could yet be ahead of us. You read it first here.