Skip navigation Scroll to top
Scroll to top

No longer behind the wheel

05 October 2017

Russell Harrop - Head of International Equities

Only a few small years ago the idea of driverless cars appearing on the roads really did seem a lifetime away. Yet in 2017 they've been tested for millions of miles in the US and are currently being tested in Milton Keynes and Greenwich in the UK with motorway testing inked in for 2019. The future is here.

The issues around the technology and its acceptance are multifarious, but broadly group into legal, ethical and practical considerations. Here's a flavour of them.

Legally who's responsible for any accident if the software's in charge? The car maker? The owner? The other party? Before getting into artificial intelligence and the ability for self-driving cars to improve their own software, they will have to be programmed with an awful lot of rules. Old school. By humans.

This leads quickly to the ethical 'trolley problem'. What to do if there's a choice between doing nothing and running someone over and doing something and potentially putting other users and/or the driver and passengers of the self-driving car at risk? It's an incredibly difficult question to answer – MIT in the US has even created a 'moral machine' website to attempt to crowd solve the problem (moralmachine.mit.edu). You're presented with various options as to who ends up being run-over and who's saved and at the end see your own (subconscious) biases. Early analysis appears to show that the range of 'answers' is incredibly large.

Practical considerations include the ability to hack a self-driving vehicle (or vehicles) at a terrifying cost to society. If there isn't such a thing as an 'unhackable car'- how would society be able to deal with this new reality? Or couldn't it?

There are plenty of benefits to self-driving technology, but by far the largest is the massive reduction in the 1.2m global deaths per annum from driving, 90% of which are estimated to be down to human error. [1] The second one is environmental. Even before considering electrification of cars (which requires a renewable grid to be truly game-changing in terms of carbon output), self-driving cars will be smoother, able to drive at the most efficient speed and be able to travel closely together on motorways making traffic jams a thing of the past. Fuel usage will go through the floor and few of us will shed a tear as lane hugging is sent to its grave. And let's not ignore the freedom self-driving cars potentially offer for the less able-bodied, the blind and our more elderly members of society, compared to today.

Despite all the hurdles, we firmly believe that driverless technology is here to stay, but given the high cost and the average car being 13 years old it's going to be some time before it's the norm rather than a 'richer person's' thing.

We find it hard to see the winners, but see plenty of losers as the vast majority of car brands become 'me-too' as who cares what name's on the boot as long as the car is comfortable, the wi-fi's good and the self-driving software's safe?

Which brings us on to one of the biggest philosophical issues: for self-driving cars to offer the best solution for society as a whole, they should have compatible software allowing them to 'talk' to each other and use system wide information to improve the performance of all cars. However, to do so further erodes the intellectual property of car companies, whilst not doing so means an arms-race among self-driving car companies to show they have the best (i.e. safest) software. It's almost a classic big vs. small government problem.

The car industry has sold the 'freedom of the open road' to drivers for a century. Compared to stop-starting and driving at 11mph across London, a car on-demand at the swipe of a finger and sitting doing whatever the heck you like whilst the self-driving car deals with London traffic is surely the real freedom? That's great for car-sharing / taxi companies, but goodbye car ownership? At least in cities?

In many ways driverless cars are at the vanguard for how society will deal with the impact of artificial intelligence and robotisation and their (likely negative) impact on jobs. We need a lot of new jobs in the UK to replace the million taxi, delivery and lorry-driving ones at risk. Oh, and in the short-term fewer accidents are great news for insurance companies, but in the longer-term those ever rising premiums are likely to be replaced by ever falling ones.

 

[1] Source: Stanford University