Can self-driving cars save lives?

9 February 2018
Part of our 'Thinking outside the box' series
Most car accidents stem from human error. But will autonomous vehicles keep us any safer? Stephen Greaves investigates.

In 2017, 1225 people needlessly lost their lives on Australia’s roads with the holiday period in particular tarnished by several high-profile tragic accidents marking a six-year spike in fatalities in New South Wales. While formal investigations uncover the causes and propose solutions, about 90 per cent of these accidents involved human error in some form. Whether this error results from having a bit too much to drink, disobeying speed and other road rules or being distracted by the kids in the back of the car, is neither here or there: surely this is too high a price to pay for simply being human?

Fast forward 20+ years to a future where human driving on public roads is a distant memory. Fleets of interconnected emissions-free driverless vehicles, summoned by an app, provide a service that safely chauffeurs workers to the office, kids to school, and families to the beach leaving occupants free to catch up on work, their favourite movie or sleep. Groceries and the latest consumer products are delivered by drones, driverless trains and trucks, while the B-doubles of 2018 are dinosaurs from a bygone age. The Australian road toll is a fraction of what it was in the 2010s, with the occasional high-profile accident due to a system glitch, an unforseen act of nature and the ever-present threat of sabotage. Driving pleasure now has to be satiated in private tracks or US styled ‘driving ranches’ that have sprung up as a profitable business in Australia’s vast open spaces.

Much has been debated about the driverless future and what we should be doing to prepare: When is it coming and in what form? Will we embrace it for ourselves and our loved ones? Will vehicle ownership and usage change? Will congestion get better or worse? Will it free up space in our cities? What are the implications for public transport, cycling and walking?

While these issues will ignite different views and opinions, it is important they do not distract the primary argument for driverless vehicles, which is safety. On-road trials in mixed traffic in the United States suggest about a nine-fold reduction in collisions where the driverless vehicle was at fault with the majority of collisions caused by human-operated vehicles. Australia has yet to embrace on-road trials, but most states are conducting controlled trials, which to date have no major safety violations to report. Sceptics are eager to point out the potential for machine failures and hacking, but the fact is driverless vehicles do not drink and drive, do not drive tired, get distracted, or use drugs, and do not speed or break laws.

So what is holding things back? First, the safety benefits of driverless vehicles are unlikely to be realised in mixed traffic, suggesting we need major infrastructure changes to accompany changes in the vehicle technology itself, which is still admittedly far from perfect.

Second, recent evidence suggests that while public opinions on driverless vehicles are moderating, more than 90 per cent of people would not feel safe in a driverless vehicle that did not have the potential for manual override, suggesting we are not close to being willing to relinquish control of the driving task to a robot. Compounding this is recent evidence from Sydney and Perth that suggests 60 per cent of people are unsure about whether driverless vehicles will make our roads safer, with 22 per cent actually thinking they will make the situation worse.

Third, while we now tolerate a considerable amount of human error on the roads, we are unlikely to be so accepting if a robot is deemed to be at fault.

Fourth, there is substantial inertia in our transport systems – Australia has one of the oldest vehicle fleets in the developed world and many fall short of current ANCAP safety standards. Consider also that various technological advancements have been made to improve safety (for example, rear and 360-degree cameras, intelligent speed adaptation), yet integration of even the most basic technologies into the general vehicle fleet has been painfully slow.

Finally, and likely to raise the hackles of some, while Australia is among the world leaders in road safety, we still by-and-large adopt the view that road accidents involving human error are somewhat inevitable and the policy focus is on their reduction as opposed to eradication. Perhaps we should take a leaf from Sweden who have a vision of zero accidents and make this a catalyst that drives our autonomous future.

Stephen Greaves is a Professor of Transport Management at the Institute of Transport and Logisitcs Studies at the University of Sydney Business School.

Related articles

03 June 2024

Funnelling congestion: How Sydney exacerbated congestion after spending tens of billions on transport infrastructure

Dr Christopher Day and Dick Day point to a need for more integrated, thoughtful planning, guided by independent research and a more accountable and transparent evaluation process.
06 May 2024

Should supply chain leaders invest in generative AI solutions?

Dr Jyotirmoyee Bhattacharjya provides three basic questions for supply chain leaders to ask to detrmine if they should invest in generative AI solutions.
02 April 2024

E-Scooters: Coming Soon to a Street Near You or Not?

Stephen Greaves (ITLS, University of Sydney) and Geoff Rose (ITS, Monash University) discuss how the proliferation of e-scooters in Australia has brought about legislative challenges, with confusing rules and safety concerns, revealing a disconnect between existing regulations and public expectations. The evolving landscape calls for a thorough examination of infrastructure, licensing, registration, and insurance to ensure the responsible and sustainable integration of e-scooters into the urban transportation system.