The uncertain ethics of driverless cars

The uncertain ethics of driverless cars

Despite some bad press of late, driverless cars, aka autonomous vehicles, are here to stay. If anything, over the next 20-30 years you can expect to see a significant uptake of autonomous vehicles across delivery, freight, commuter and personal transport networks. Autonomous vehicles are set to revolutionise the way we travel, yet there are still some important ethical dilemmas to be resolved.

We spoke with Professor Ani Desai, Director of La Trobe University’s Centre for Technology Infusion, about what’s needed to make self-driving cars the transport of the future.

driverless cars, accident, vehcile, ethics
Driverless cars need to react in real time. Image credit: Unsplash

How do driverless cars work?

In the same way that our eyes, ears and brain help us drive a car, an autonomous vehicle takes inputs from a range of sensors designed to mimic human perception. The autonomous vehicle’s sensors are integrated through a central control system, like a brain, that interprets the data and controls when the vehicle stops and when it moves.

In the future, autonomous vehicle technology may become sophisticated enough to all but eliminate accidents.

But in cases where accidents are unavoidable, the ethics of autonomous vehicles become murky.

This is because each autonomous vehicle makes decisions based on pre-programmed logic that may be unable to deal with complex ethical situations.

To illustrate, Professor Desai offers this example:

Imagine you are driving down the Great Ocean Road in your driverless car on a busy, sunny Saturday. Coming round a bend you see people standing all over the road, oblivious to your approach. Without enough time to brake, does the driverless transport:

  1. Run you over the cliff – endangering you but definitely saving all of the people on the road?
  2. Run into the crowd – saving you but possibly injuring several people?
  3. Run into oncoming traffic – taking a low percentage chance to avoid all injuries with the chance that the outcome will be the worst of all three scenarios?

According to Professor Desai, the issue is not what autonomous vehicles can do – in theory, they can do anything – but what we will allow them to do. At this moment in time, regulators and industry groups in Australia are a long way from a coherent set of rules for autonomous vehicles in such challenging situations.

What’s needed to regulate self-driving cars?

Professor Desai suggests that a multi-disciplinary panel of experts is needed to create regulatory frameworks for how autonomous vehicles should act in extraordinary situations.

‘If those ethical frameworks are not there, how will you place accountability for automated decision making? Does the fault lie with the company, the software writer or the owner?’

On people’s immediate fears around the safety of autonomous vehicles,

Professor Desai says the novelty of an autonomous vehicle accident means incidents involving self-driving vehicles get overblown.

‘When there is a death created by an autonomous vehicle, it is reported everywhere, even if it was not the fault of the machine itself.’ Human to human car accidents (except for the most horrific) are almost invisible by comparison.

car, driverless, self-driving
A lot of future transport might be autonomous. Image credit: Unsplash

The future of transport

Despite a few teething problems for driverless cars, Professor Desai says we can expect them to feature prominently in the future of transport. Transport that is ‘electric, automated, connected and shared.’

La Trobe is making its own contribution to the future of transport by trialling the autonomous shuttle bus, a Victorian first, at the university’s Bundoora campus. Professor Desai says that one of the first mainstream uses for autonomous vehicles will be in ‘last mile connectivity’ – ferrying passengers between transport hubs and their destinations. Keep an eye out, soon you might be catching a driverless bus from the train station to your lecture theatre.

Want to help build the future of autonomous transport? Study a Master of Data Science or a Bachelor of Engineering Honours (Industrial) at La Trobe.