Imagine this scenario of the not-too-distant future: as the driver, you have handed control over to your autonomous car and are heading down the Tuggeranong Parkway, the software perfectly matching maximum posted speeds and distances with the cars around you in what should be a safe, stress-free and expedient journey home.
But suddenly a kangaroo bounds out of the underbrush beside the roadway and in front of the car ahead of yours, which automatically executes an emergency braking manoeuvre.
This triggers a similar response from your car, and all the vehicles around. Some respond faster and than others, with perhaps not-so-adept software or dud brakes, respond slower. The driver of the 15-year-old Volvo behind you and which doesn't have an autonomous system, responds even slower again.
A huge crash ensues, multiple expensive cars are badly damaged and people injured. From an insurance perspective, who or what is at fault?
This, and various other real-life scenarios like it, creates conundrums to which there are no easy answers.
Who is at fault when the system fails: is it the driver, the car itself, the external systems which the car needs, or the company which built the car?
The National Transport Commission, the peak advisory body to federal, state and territory governments on transport reform, is wrestling with this issue to develop a framework for the safe deployment of automated vehicles on our roads.
The car industry wants to shift the blame for the crash away from the automated car, and says this will require "a new state and territory policing philosophy".
"Where [automated] vehicles contravene road rules, the focus should not be punitive, it needs to be focused on the primary causal factors; this will require a complete rethink of the way local jurisdiction policing is undertaken," the car industry lobby said.
It is also critical of not having the broad telecommunications coverage across the country which automated cars would need for geo-location and control functions.
"The [discussion] paper is currently silent on the critical role of the telecommunications industry," it said.
The transport commission has acknowledged that there is still much difficult ground to cover in the discussion while various trials continue around the country.
"Existing regulation and legal frameworks developed for regulating conventional vehicles and drivers do not adequately cover new risks that are unique to automated vehicles," it said in its summary.
All the key stakeholders involved are wary of the legal issues and entanglements which would certainly arise if every jurisdiction was given the decision to roll out its own set of rules, as currently exists with driver licensing and registration.
So what is proposed is a single national regulator - presumably sitting within the federal Department of Transport - to manage "in-service" automated cars; that is, those cars that will be driving around on our roads within the next 5 to 10 years, and an Automated Vehicle Safety Law (AVSL).
The proposal is that for every automated driving software system - known as an automated driving system entity - there will be self-certification against a set of safety criteria and obligations before they are granted a type approval to enter the market.
And that entity will need to have what is known as a principles-based "general safety duty" under Commonwealth law. That primary duty is to keep the occupants safe.
The insurance industry, which is understandably jittery about tenuous legal frameworks - who carries the cost of car repairs and more importantly, personal injury liability when a crash happens - strongly supports a centralised plan.
