Self-driving cars claim to be safer but the world lacks evidence to prove it
A new type of car technology is said to be up to 90 per cent safer than current vehicles but are its promises too good to be true?
OPINION: Australia is set to expect self-driving technology any day now, but the truth is we have no idea if it will actually make our roads safer.
We remain stuck in a grey zone, with no national crash-reporting system for driver-assistance or autonomous vehicles, and no clear picture of whether these systems are truly safer than human drivers.
For now, we’re taking automakers at their word.
Tesla reports its cars are almost 10 times safer with its driver-assist software engaged, while Google-owned Waymo and an independent study from the University of Toronto suggest its driverless cars crash up to 60-90 per cent less often than human-driven vehicles.
But in Australia, where over 1,300 people died on the roads in 2024, there is no local data to verify these claims.
MORE: Australian EV disaster looms
We know human error is to blame for the majority of fatal crashes – distraction, fatigue, speeding, and alcohol.
On paper, automation should help eliminate these problems. Computers don’t get tired or drunk.
FASTER THAN HUMANS
They react faster than humans, who take an average of 1.45 seconds to respond.
But if we look at what’s happening overseas, it is far from reassuring.
Tesla’s Full Self-Driving has been linked to hundreds of crashes and at least 44 confirmed deaths in the United States, according to the National Highway Traffic Safety Administration (NHTSA).
A 2024 investigation by the NHTSA identified a “critical safety gap” in autopilot tied to 467 crashes.
Common complaints include unintended acceleration, sudden braking, and vehicles veering off highways.
RELATED: Why Audi’s RS 3 represents the end of an era
‘MONITORING PARADOX’
Then you have the “monitoring paradox” when drivers disengage, expecting the car to do the work and then fail to react when the system falters.
But machines have their own limitations.
These systems have struggled with “edge cases” such as construction zones, unusual intersections, or poor weather. Critics point to a “monitoring paradox,” where drivers may mentally disengage and then fail to react quickly when the system falters.
MORE: China fires back in EV speed war
Without a national reporting scheme in Australia, a crash involving a self-driving system is currently logged as human error because the driver holds all legal liability.
The Federal government is preparing new legislation, the Automated Vehicle Safety Law, which is expected to shift some responsibility from drivers to system providers.
Advocates and safety bodies say Australia must establish a mandatory crash reporting hub to assess these systems against the reality of human drivers.
Until then, it remains impossible to know if they truly improve safety on Australian roads.
Originally published as Self-driving cars claim to be safer but the world lacks evidence to prove it
