It's the thorny ethical dilemma facing developers of self-driving cars: who to kill.
As Volvo pushes for a trial of self-driving cars on Sydney roads, new research suggests that people generally approve of self-driving cars sacrificing their passengers to save others – but they're not so keen to ride in the utilitarian cars themselves.
This conflict between moral values and personal self-interest shows just how difficult it will be to make decisions about programming the cars, despite expectations the vehicles will dramatically improve road safety.
As part of the research, published in Science on Friday, Jean-François Bonnefon of the University of Toulouse led six online surveys of US residents between June and December 2015 and asked participants how they would want their driverless cars to behave.
Respondents did not think cars should sacrifice a passenger when only one pedestrian could be saved, but their approval rose with the number of lives that could be spared. When asked to imagine a family member in the car, their approval dropped.
But while participants approved overall of cars that sacrificed passengers for the lives of 10 others, they were less likely to buy the cars themselves.
Research leader Professor Toby Walsh, of National ICT, said it was an interesting response given people had already accepted a lack of control elsewhere, such as on planes and trains.
"We will undoubtedly get to that point here as well, and I suspect in 20 or 30 years' time we'll look back and think: 'it was the wild west on the roads back in the day of Grandpa – we allowed people to drive such lethal weapons around'," he said.
Professor Walsh said that given the survey was based on US residents, it was difficult to speculate what the research would mean for Australia. However he said he suspected public sentiment would be similar.
The paper said one typical solution to the dilemma would be for regulators to enforce the behaviour that would lead to the best global outcome, offering immunisation laws as an example.
However, Dr Bonnefon, the study's author, said researchers were surprised by the participants' responses.
"If driverless cars were utilitarian by law, people would continue to drive themselves as long as possible, forfeiting all the safety benefits of driverless cars out of fear that their own car might one day decide to kill them – a very unlikely occurrence," he said.
The results suggest regulation could delay the adoption of the cars, something the paper and Professor Walsh said could pose an ethical issue in itself.
"We really need to solve this problem because we want to get people in autonomous cars to make the roads so much safer," he said.
"There's a moral imperative in the sense that 1000 people will die on the roads of Australia over the next year and 95 per cent of those accidents are driver error.
"If we can remove the driver, we can reduce those statistics down to very small numbers."
The study said ethical challenges had to be addressed and rules made before driverless cars became a global commodity.
It said manufacturers and regulators would need to achieve three potentially incompatible objectives: being consistent, not causing outrage, and not discouraging buyers.
Professor Walsh said the cars were already part of our lives, and he expected that in a decade's time the majority of cars sold would be autonomous.
'Laws and even our societal acceptance is yet to catch up with where the technology is," he said.
Earlier this month, Drive reported Volvo's submission to an inquiry that self-driving cars would be on sale in Australia by 2020.
The Swedish brand called on the NSW government to trial the cars on Sydney roads within the next five years.
Fairfax Media
