STRICTLY EMBARGOED UNTIL 00.01 ON FRIDAY 14TH SEPTEMBER 2018
- Motorists ‘programme’ a driverless car in new experiment -
- Moral dilemma: New report1reveals the decisions drivers think an autonomous vehicle should make on their behalf –
- The Conscious Car simulation2 allows drivers to programme the moral driving decisions of an autonomous vehicle (AV) and compare their results against the rest of the UK.
- The majority (58%) of ‘programmers’ protected pedestrians over car occupants and were more likely to save children (59%) over the elderly (9%).2
- EXPERT REPORT1 provides clarity on the transitional challenges for the transport network and driver safety, despite UK Government plans to have AVs on the road by 2021.
- Report discusses whether funding for AVs should be devolved and asks if they will solve socio problems or inflate them.
Drivers have been asked to ‘programme’ an autonomous car with moral driving decisions as part of a revealing experiment.
The Conscious Car simulation2 launched by Confused.com, allows users to programme an autonomous car by asking them to make a series of moral driving decisions on its behalf. The simulation presents ‘programmers’ with a number of dual-choice scenarios whereby they have to choose the vehicle’s action in the event of an unavoidable accident. For example, users are asked to choose between protecting the occupant of the car vs an adult pedestrian. And, perhaps more controversially, one of the scenarios requires the programmer to choose between protecting a child vs an elderly pedestrian.
The data collected from the experiment provides clarity on the how driving population think autonomous vehicles should act on their behalf. Based on majority responses across different scenarios2, overall drivers felt autonomous vehicles should be programmed in the following way:
- Prioritise pedestrians over occupants (58% vs 27%).
- Take a neutral position on the number of lives saved (46%).
- Protect humans over dogs (84% vs 16%).
- Protect children (59%) over the elderly (9%).
The experiment forms part of a new expert report1 into autonomous vehicles. Following the Automated and Electric Vehicles Act 2018, which received royal assent in July this year, the report explores whether the new technology is a ‘driverless dream or a nightmare in waiting’. It provides clarity on a range of confusing questions about driverless vehicles, for example, whether the UK Government’s ambition to have autonomous vehicles on our roads by 2021 is unrealistic.
The report also discusses some of the transitional challenges associated with autonomous vehicles, such as improvements to road markings in order for the optical sensing technology of AVs to be able to work effectively. And the potential impact on road safety if vehicles with varying levels of autonomy are on the road at the same time. For example, motorists have been found to drive ‘sub-optimally’ for up to 55 seconds after taking back control of their car following an autonomous driving period. And the over-cautious nature of the first self-driving cars could lead to a slowdown in the pace of the road network and increase the number of accidents.
The report also explores the challenges of funding infrastructure to support the new technology. For example, if funding for AV infrastructure was devolved to local authorities there may be a disparity in prioritisation, which could lead to a breakdown in the system for drivers travelling across counties. And at a philosophical level, the report discusses whether autonomous cars have a place in solving socio problems, such as social isolation and congestion, or whether they inflate them.
Finally, the report goes into detail about the programming of autonomous vehicles in the event of an accident. And it’s clear from the above findings that not all drivers have the same opinion when it comes to how to autonomous vehicles should prioritise protection.
Sample of driver-decisions for scenarios in The Conscious Car simulation2
The research reveals a clear difference in the moral programming made by certain demographics, too. For example, while both men and women were most likely to programme the vehicle to protect pedestrians, women were more likely to make this empathetic decision than men. While men had an increased tendency to save the occupant of the car. Women were also more likely than men to programme the car to protect multiple people and dogs. However, men were more inclined to make decisions to protect the elderly than women.
Comparison of programming decision weighting between men and women:
Most likely to protect…2
There was also variation in the programming decisions made depending on the age of the programmer. When compared to other age groups, younger drivers (18-24) were more likely to take a liberal stance and programme their vehicle to protect pedestrians and many individuals over the few. While older drivers, particularly those aged between 45-54, were more likely to protect the occupant of the car. Perhaps this is indicative of this age group having an increased sense of personal welfare, as this age range is when drivers are most likely to become grandparents for the first time3. Younger drivers were also more likely to have a soft-spot for animals and programme their vehicle to protect dogs, compared to other demographics. While drivers over the age of 55+ were least likely to choose to protect the elderly, in what some might describe as a selfless act.
Comparison of programming decision weighting between age groups:
Most likely to protect…2
Given the findings of the report and experiment, it’s no wonder drivers seem to be split on whether they would use an autonomous car. While 46% said they would, a third (32%) said they wouldn’t and a further one in five (22%) weren’t sure either way.3 There are clearly a number of challenges to overcome before the technology can be fully integrated across our transport network.
Amanda Stretton, motoring editor at Confused.com, says: “Our report makes it clear society needs to have some serious debates before driverless technology is introduced to our roads.
“The decisions driverless vehicles make will ultimately be decided by humans at the programming stage, so it’s important to start having conversations about how we want this technology to think and operate.
“Drivers who want to see where they stand compared to the rest of the UK can try programming an autonomous vehicle using our Conscious Car simulation.”
Notes to editors
- Confused.com’s report was published September 2018. It collates research from a number of different sources (referenced in the report). Featuring contributions from Matthew Avery, Head of Research at Thatcham Research, Stu McInroy, Chief Executive at Road Safety Markings, and Dr Ian Walker, Environmental Psychologist at Bath University.
- All data obtained from Confused.com’s Conscious Car Simulation based on 2,000 submissions in August 2018.
- Average age of a first time grandparent: https://www.telegraph.co.uk/news/uknews/1494949/Average-age-of-the-first-time-grandparent-is-now-under-50.html
- Nationally representative survey research to 2,000 UK drivers in August 2018.