Imagine 2026 – just 3 years from now – you have just taken delivery of your new car. It drives itself, you don’t have to steer, brake or anything. Australian governments are planning to have the necessary law in place by 2026 to allow you to drive your new autonomous car anywhere in Australia. This article discusses the potential roadblocks between now and the expected 2026 legislation, the legal and insurance developments and where Australia is compared to the rest of the world in preparing for autonomous vehicles.
The road for autonomous vehicles is bumpy
The 2020 KPMG Autonomous Vehicles Readiness Index had Australia ranked 15 out of 30 countries reviewed for overall readiness.
I could not find a 2022 update. There may be a reason for that.
If you go to the Federal Governments’ National Transport Commission (NTC) website on the preparation for autonomous vehicles and click a link to take you to current reforms here, it refers you to the September 2020 Automated Vehicle Program Approach.
One of the NTC pages relevant to CTP insured injury law reform had the following:
“Transport ministers agreed in August 2019 on a national approach to insurance for automated vehicles that requires existing motor accident injury insurance schemes to provide cover for automated vehicle crash injuries.
Ministers responsible for motor accident injury insurance schemes, through the Board of Treasurers, have been asked to endorse the transport ministers’ decision. A treasury officials’ working group is considering the policy issues and options identified in this project and will report to the Board on progress”.
I could find no reference to any progress report from the Board of Treasurers (which excludes from the Board the Federal Treasurer). The Transport Ministers did meet in February 2022 (the ITMM) to again consider this. The NTC also published in February 2022 a policy paper: The regulatory framework for automated vehicles in Australia.
That policy paper, relevant to CTP, had a section which referred to a 2019 NTC document, Motor accident injury insurance and automated vehicles, which was agreed by the ITMM in August 2019 as referred to above.
This is a bit of a circle which seems to suggest nothing has been done on CTP and other law reform for autonomous vehicles since 2019 in Australia. That is not quite true. Something is being planned for 2026 (see below). To be fair to the NTC it has a lot on its plate and autonomous vehicles are just a part of its brief. The same can be said for the Treasurers and Transport Ministers.
Heading in the right direction
The 11 February 2022 ITMM communique indicated agreement that the Commonwealth would legislate an ‘Automated Safety Law’, to ensure consistent regulation across the country for everything to do with autonomous vehicles. This was to have an intergovernmental agreement to support the law reached by “late 2023”. The NTC, the Commonwealth and State governments would work to develop complementary legislation and report back to the ITMM by the end of 2022. There was nothing about this in the ITMM communique for their 9 December 2022 meeting. The commencement of the Automated Safety Law, expected by 2026, assumes that deadlines, like the December 2022 one, are met.
Without being exhaustive as to things happening in Australia, another area of progress is Australian State and Federal governments wanting to support Cooperative Intelligent Transport Systems (C-ITS). These are interconnected systems to allow vehicles to communicate with other vehicles, road infrastructure, data services and other road users including pedestrians and cyclists. Once established, C-ITS ideally not only would enhance the performance of automated vehicles but also support other vehicles once adapted. Public consultation on principles underpinning this development in Australia closed on 3 February 2023.
There are many pilot projects, an excellent example of which is the Ipswich Connected Vehicle Project (ICVP). This wrapped up with final reporting in May 2022 after testing throughout 2020 and 2021. It applied C-ITS retrofitted to vehicles with more than 350 participants, 2.7 million km travelled and 49,000 driving hours. As the report found, the benefits of C-ITS are clear:
“In a broader context, with a total of 526 fatalities and 20,826 serious injuries being reported from crashes on the South East Queensland (SEQ) road network over the five-year period of 1 July 2016 and 30 June 2021, the eight C-ITS use cases implemented in the ICVP could have prevented up to 101 fatalities and 4,198 serious injuries from crashes. This equates to an average of 20 fatalities and 840 serious injuries prevented each year in SEQ”.
The NTC is also researching the question of law enforcement for automated vehicles and between July-September 2022 sought public comment to be able to report to the relevant Transport Ministers. This was not in time to help Victorian police with the defence raised in March 2022 by a person charged with alleged traffic offences when the person driving a Tesla, claimed to be in ‘autopilot’ mode, injured a pedestrian. This mode still required active driver supervision, is not autonomous and the driver must keep their hands on the steering wheel. The outcome of this defence is not publicly available.
Where do we stand?
The question is where Australia stands relative to those countries leading the advance of autonomous (and electric) vehicles? The agencies and Ministers in those countries also have a lot on their plate but seem to be making significant advances:
“Germany has also offered comprehensive legislation on AD [autonomous driving] that has allowed one European OEM [original equipment manufacturer] to launch the first true L3 feature in a current model. Similar legislation exists in Japan and has recently been authorised in France. The development of global AD standards for private-passenger vehicles is clearly in motion”
- Autonomous driving’s future: Convenient and Connected, McKinsey.com January 2023.
L3 (Level 3) means an autonomous vehicle which drives itself and the driver does not need to watch the road but must respond and regain control (by manual steering, speed, braking etc.) when the vehicle requests. The levels go from 0 (no assistance, not even cruise control) to 5 (fully autonomous, e.g. no steering wheel). The OEM in this example is Mercedes-Benz and the L3 is currently available in Germany and can be driven on public roads. The law to enable the driving of autonomous vehicles in Germany was first enacted on 21 June 2017 and then on 28 July 2021, the Act Amending the Road Traffic Act and the Compulsory Insurance Act (Autonomous Driving Act) came into force enabling the L3 Mercedes-Benz and vehicles like it to drive in Germany. By 2024 the EQS Sedan and S-Class will be the first models released to the US with L3 autonomous capability with testing in Nevada. There is much discussion surrounding Tesla’s autonomous vehicles and this paper is too short to do this justice. Suffice it to say that Germany and the US are very advanced in this field, as is China.
The UK is sometimes a source for Australian legislative development. The Automated and Electric Vehicles Act 2018 (UK) already covers the liability for autonomous vehicle accidents. The legislation is relatively brief and is almost there as a precaution pending more thorough legislation. In a joint report in January 2022 by the Law Commission of England and Wales and the Scottish Law Commission a comprehensive analysis was done regarding the necessary regulation. This was based on broad stakeholder input which started in 2018. The recommendation was to introduce a new Automated Vehicles Act including a test the vehicle must satisfy before being authorised as having a self-driving ADS feature (automated driving system). An in-use regulatory authority was recommended to be established (including having the power to collect data). Safety was a focus including addressing that in the transition from assisted driving to fully autonomous vehicles, there is a risk that drivers could be misled by marketing into thinking that they need not pay attention when their vehicle is not yet fully autonomous.
Despite marketing enthusiasm, fully autonomous vehicles cannot legally be driven by the public and initially will not be easily affordable – there is not yet public demand. Safety concerns exist. The vehicles require more sophisticated guidance systems than aircraft because of the complexity of roadway interactions1.
In fact the drive to further fund and develop the necessary technology apparently has slowed. In recent months Argo AI, an autonomous vehicle company backed by Ford and Volkswagen, was shut down. The investment by Ford was shifted from trying to develop L4 autonomous systems to more achievable L2 and L3 technology within existing Ford production. Volkswagen shifted its investment from Argo AI to existing partnerships with Bosch and Horizon Robotics in China. Motional (a joint venture between Hyundai and Aptiv) “announced steps to reallocate resources that will help ensure long-term commercial success” according to a company statement in response to a rumour of staff lay-offs late in 2022. This is the same company that in October 2021 was exploring with the New South Wales government to pilot the use of robotaxis in NSW. When I searched for more detail on this pilot all I could find was the October 2021 announcement.
Although some trimming is happening, there continues to be massive investment in the development of autonomous vehicles. Soon, it is inevitable that at least in some countries there will be laws and insurance underpinning increasing use by the public at least up to L3 vehicles, like those Germany already has in place. The progress to L5 fully autonomous vehicles being broadly available to the public, however, likely will take decades and investors want returns much sooner than that.
Predictions for Australia
Will Australia have widely available L3 autonomous private vehicles and up to L5 public transport by 2030? These will need to be governed by the promised Federal Automotive Safety Law 2026 with supporting cooperative legislation from the States. This law will need to cover everything that we currently take for granted with conventional vehicles but tailored to autonomous vehicles. It will need to regulate, for example, what a police officer can do with the data from your vehicle when they pull you over for alleged speeding. Existing product liability law is cumbersome for motor vehicle accidents, so the new law must define what needs to be proved to establish liability of the company that, for example, supplied/programmed the component which failed and caused the vehicle to catastrophically injure a pedestrian. New road rules will need to apply – where and how an autonomous vehicle can be used.
Todd Litman of the Victoria Transport Institute has bravely attempted to make some predictions. In his article, Autonomous Vehicle Predictions – Implications for Transport Planning (25 January 2023 https://www.vtpi.org/avip.pdf), a main limitation to the technology identified is the expense of privately owned autonomous vehicles. Until the technology becomes more affordable - Litman identified that this is probably the 2040s to the 2060s - the full benefits will not be gained. As McKinsey research has suggested:
“Steep up-front costs for developing L3 and L4 driving systems suggest that auto companies’ efforts to commercialise more advanced AD systems may first be limited to premium-vehicle segments”
Insurance coverage for autonomous vehicles is a related issue to the need for law reform. The interesting McKinsey interview of Andrew Rose (OnStar Insurance and Executive VP of GM Financial), ‘Future vehicles require future insurance’ covers the need for insurance to adapt to both electric and autonomous vehicles. Whilst the article addressed US and global insurance issues, this is highly relevant to Australia. As with all insurance, data remains key, including to segment and price products, reward drivers for good performance, manage claims based on real time data from accidents, including identifying liability. Information from sensors, including video recording and onboard data recorders built into vehicles increasingly will be available to insurers. Much of this has already been available but the collection and use of the data will become more and more sophisticated. Privacy is an issue here as well.
Adequate infrastructure is one major goal Australia needs to achieve for autonomous vehicles to be publicly available and why the C-ITS investigation referred to above is critical. For example, designated autonomous vehicle lanes may exist in the transition period until L3-L5 vehicles become more available. This may mean one lane for the wealthy who can afford the autonomous vehicles and one for the rest of us.
More likely is that rather than the ‘fast lane’ being used by the wealthy, it will be filled with autonomous buses and taxis. ‘Robotaxis’ may become the popular and accessible option for autonomous transport. In 2021 Baidu was testing them in Beijing and Waymo trialled them in San Francisco. In 2022 GM’s Cruise made them publicly available in San Francisco. Motional was trialling their robotaxi in the US and Singapore in 2022 and, as noted above, apparently are in discussions with the NSW Government.
Imagine Australia in 2026 and a millionaire sitting in their jacuzzi getting quietly sozzled on the best wine from Maclaren Vale in the cabin of an advanced L5 autonomous vehicle – with no steering wheel at all. Or a party in a robotaxi – again no driver – driving down the designated lane. What if the vehicles collide with each other or a pedestrian? Imagine if the promised 2026 Australian national law never happened and was still bogged down in a discussion paper or parliamentary committee. The law as it is in 2023 in Australia cannot adequately deal with this. In Germany the law is already up to speed.
Australia is being left behind
Many vehicle collisions are caused by human driver error. There is much disagreement as to what is the percentage. Regardless, imagine how this can be reduced significantly by the safety advantages of autonomous vehicles. Of course there will still be accidents caused by autonomous vehicles but fewer than those currently caused by human error. Imagine our emergency departments not clogged by road trauma. There will be more accessible transport for those with disability. Environmental and productivity benefits flow from more efficient transport. These are some of the reasons why other countries have been investing and continue to invest in this technology, the law and infrastructure behind it.
At the October 2021 summit on driverless vehicles presented by the then peak Australian and NZ industry body ADVI (Aust. & NZ Driverless Vehicles Initiative), I and other presenters, including from Singapore, the US and the UK, addressed where we were then and what the future held. The conclusion was that Australia needed to do more. I would say now that Australia is well behind, not even in the rear-view mirror of the L3 Mercedes S Class driving in Germany.
ADVI was essentially a volunteer organisation which was effective at sharing ideas on autonomous vehicles between those in government, industry, academia and the broader community. It finished in 2022 with a demonstration of autonomous test vehicles at the NSW Government autonomous vehicle test site at Cudal. ADVI was replaced with a different body, espousing essentially the same objectives and with a catchy acronym, CCAT - The Centre for Connected and Automated Transport. CCAT was launched in late October 2022 and the hope is that the next time I write, CCAT, or for that matter the NTC or the Transport Ministers – or someone – can demonstrate more rapid progress in Australia for autonomous vehicles. Even if the technological development is slowing to further ensure safety and build consumer confidence, Australian law is still several years behind where it needs to be.
Australia needs to catch up.