Driverless cars taking us nowhere

Brum, the first driverless car?

Driverless cars are never going to become a reality, let’s face it.  If we all get off the bandwagon for just a moment, step back a few paces and look from afar, that is only conclusion which we can come to.  In the meantime, we’re wasting time, money and the brainpower of some very clever individuals by pursuing this mirage.

The reasonable man test

In English law, there’s a concept promoted by the deceased Lord Denning about what the reasonable person might do in a given situation.  He called this fictitious individual the man on the Clapham omnibus.

Driverless cars are a great concept.  The reasonable person approach allows a case to be tried not on the basis of attempting to prove what the individual in a particular situation felt, but on what an ordinary person, acting reasonably in that situation might have done.  We then compare that to what the individual actually did do, and we have an approach to liability.  (I’m not a lawyer, don’t run over me with your driverless car.)

Though a useful concept, it’s often been criticised because despite all logic, people often don’t act reasonably.  Many of us think that acting a little unreasonably is completely reasonable, a tautology, if ever there was one.

What does all this have to do with driverless cars?

It’s one thing having a car drive down a three or more lane highway, or to follow a road, or park a car.  It’s quite another to trust that technology with little Jimmy’s life when he runs into the road to collect his football.

Of course, we could build the technology to stop the car automatically.  In fact, it exists, works well and has been used on our roads already.

Let’s reshape little Jimmy’s situation.  Let’s say Jimmy and his friend Andy both run into the road to collect their precious ball.  Now let’s throw both a man pushing a buggy and woman carrying a heavy load of shopping along the pavement.  In other words, a pretty normal street scene.

Now when our little friends run into the road, our driverless car has to try to:

  1. stop suddenly, ,the preferred option, or
  2. swerve out of the way.

Our algorithm gets to work immediately, only to discover that the first option cannot be completed in time without hitting one or both of the children.  Clearly, this is an inadequate conclusion to this incident.

Next, our algorithm moves on to the second option, swerving out of the way.  In this situation, the swerve/ break combination might still involve one of these five people, Jimmy, Andy, the woman with her shopping or the man with the buggy and the baby inside.

Who should the driverless car choose to hit, whose life is more valuable?

Should the baby go, because it’s the youngest (in English taut law, this might be one approach to minimising damages).  Should the driverless car go for the woman with the shopping, despite the fact that she stuck to the rules of the road?  Perhaps it should go for the kids, punishing them because they broke those very same rules?  In that case, which kid should the car hit?

Should the driverless car hit a child, despite being likely to die as a result, or should it hit the innocent man, just because he’s slightly less likely, leaving a baby fatherless?

Would you trust an algorithm to make those decision?

We often aren’t even that keen on the product recommendation algorithms, or the Google ad ones.  I despair when Cortana or Siri starts calling mum when I’d said “set a reminder at one”.  We struggle to make a decent responsive website, failing faster and faster but without much success.

That’s why we have people

That’s why, of course, all driverless vehicles come equipped with somebody who can take control in an emergency, at a touch of the steering wheel, for example.

We trust the people who drive cars (to the extent that we trust anyone, anyway).  In these situations, the man on the Clapham omnibus often isn’t the right comparison for us and the decisions we make.  He makes reasonable decisions, whereas individuals, each and every one of us make complex, sometimes undefinably right, decisions.

The decisions which we make, based on all our experiences, often can’t be explained to the man on the Clapham Omnibus, let alone made by him.

And after all that, if we are still required to put a driver in the car, what progress have we actually made?

Driverless car technology is dead, long live driverless car technology

Perhaps the future of transport is more likely to be something like Elon Musk’s Hypersphere.  Musk has planned the first one, a little tube which is planned to whisk people between cities at the speed of sound, to connect Los Angeles to San Francisco, minimising the travel time to half an hour.

Whatever the future holds, it doesn’t mean that we’re completely wasting our time with driverless technology, merely that we should be more realistic about the ambitions and uses for it.

Helping people drive and sensors to improve detection are great to aid safety.  These technologies are becoming a welcome reality.

Financial incentives

One place to look at whether, and how, a business venture, from purple spaghetti to driverless cars is going to become a reality is the financial value.  Who benefits the most from the technology?

It’s hard to see how the individual driver is the group most likely to benefit from driverless vehicles. They still needs to be able to drive, they still needs to pay attention to the road so that they can control the car.  They still need to take driving lessons and pay for driving tests.  They’re still, presumably, responsible in the event of an accident and paying a healthy insurance premium too.

It’s more likely to be businesses who would benefit, in particular, those organisations which haul huge loads of goods around our roads.

To these businesses, technology for driverless cars might represent higher safety, which helps them to reduce costs.  Potentially, it might result in cheaper labour.  It could also help them predict driving times along key routes, helping them manage their operations and reduce costs there too.

What does tomorrow hold?

Overall then, driverless cars and other vehicles are something to look forward to.  It’s just that when they come, we’ll still need drivers.

Isn’t it time to look for the true technology of tomorrow?