A hallmark of the twenty-first century artificial intelligence renaissance is the autonomous vehicle (AV), an automotive advancement that has pushed scientific frontiers and is poised to transform the transport industry. Though still in their infancy, it is conceivable that AVs will set the wheels in motion for a mobility revolution and will pave the road to increased productivity, efficiency and safety. However, there are significant liability risks and uncertainties associated with AVs, particularly in relation to collision liability, the Queensland Road Rules, and use of data created by AVs. To ameliorate uncertainty, Queensland should implement a no-fault insurance system and amend the law to resolve AV data ownership issues and clarify liability under the Road Rules.
AUTONOMOUS VEHICLES
A The Technology of AVs
AVs are an example of weak artificial intelligence[1] characterised by autonomy, reactivity, goal-centeredness and temporal continuousness.[2] An AV’s computer processes input via cameras, GPS, radar and lidar in order to understand the AV’s surrounding environment and make driving decisions.[3] This is achieved through machine-learning algorithms including deep neural networks, convolution neural networks, regression, pattern recognition, clustering and decision matrix algorithms.[4] Generally, vehicles are classified under one of five levels of automation ranging from no automation or conditional automation to full automation.[5] Buses, trucks, emergency vehicles and planes are all potential types of AVs,[6] but self-driving cars provide the greatest scope for discussion.
B Benefits of AVs
AVs are significantly less susceptible to everyday human driving errors[7] and may therefore reduce road fatalities, injuries and lawsuits associated with negligent driving.[8] Furthermore, AVs offer the potential to alleviate traffic congestion[9] and maximise fuel efficiency.[10] For example, one study found that self-driving cars could improve highway capacity by up to 273%.[11] AVs may also provide social and economic benefits: 70% of Australians support self-driving cars,[12] and it is predicted that AVs will save the US economy $1.3 trillion per year through productivity gains and accident avoidance savings.[13]
C Why Consider AV Liability Risks?
It is imperative that vehicle users, manufacturers and other interested parties are acutely aware of the liability risks associated with AVs. These risks are starkly illustrated by some recent collision incidents. On 7 December 2017, a General Motors self-driving car collided with a motorcyclist who was attempting to overtake it, resulting in a lawsuit.[14] In March 2018, an Uber self-driving car in autonomous mode killed a woman crossing a dark road in Arizona.[15] These incidents raise significant questions regarding the imposition of criminal liability upon AV users and manufacturers[16] and the appropriate negligence and product defect standards that should apply.[17]
In addition, the Australian Competition and Consumer Commission has been steadfast in its commitment to the enforcement of product safety and consumer guarantees in the automotive industry.[18] The Product Safety website has recorded over 3900 transport product safety recalls in its history, and there has been a consistent year-by-year increase in recalls of this type over the past decade.[19] Recently, the automotive industry has also faced scrutiny in the wake of a mandatory Takata airbag recall.[20] Accordingly, retailers and manufacturers of AVs may face substantial risks of prosecution if they do not comply with mandatory product safety standards or consumer guarantees.
COLLISION LIABILITY
A Attribution of Liability
A key issue to determine in evaluating collision liability risks is who will be held predominantly responsible for an accident. This warrants a consideration of competing policy arguments in relation to the attribution of liability to relevant parties.[21]
1 AVs as Legal Agents
Many academics have postulated that AVs can be considered genuine legal agents and can therefore be held legally responsible for their own actions.[22] This contention is grounded in utilitarian concerns and is predicated upon the notion that no metaphysical barriers restrict the attribution of agency to autonomous machines.[23] However, the law does not currently afford legal agency to AVs and to impose liability would ‘venture into new ground.’[24] Moreover, it may be difficult conceptually to regard AVs as legal agents,[25] and full legal autonomy arguably requires moral choice.[26] Therefore, it is improbable that AVs will be held liable as legal entities in their own right in the foreseeable future.
2 Manufacturers and Users
More feasibly, attribution of liability issues will turn upon the tension between holding manufacturers responsible for crashes and imposing strict or fault-based liability upon AV users. On the one hand, it could be argued that car manufacturers should be held mainly responsible for any crash caused by an autonomous vehicle that they designed.[27] This would ensure that manufacturers do not economise on safety and is consistent with two overarching goals of ‘liability law’, namely minimalising accidents and compensating victims.[28] In a press release, vehicle manufacturer Volvo supported this position, stating that it would accept full liability for its cars when they are in autonomous mode.[29]
However, on the other hand, many academics posit that imposing an excessive liability burden on manufacturers may stifle innovation and technological advancement.[30] On this basis, it is contended that AV users should observe a duty to pay attention,[31] returning liability to the driver where they had the ability to intervene to prevent the accident.[32] Some academics even suggest that AV users should be strictly liable because they assumed the risk of using an AV.[33] If taken too far, however, imposing liability upon users of AVs would be equally problematic. Morally, it may be a form of defamation to blame a person for their inattention if they had no real chance to intervene.[34] Further, manufacturers should not be disincentivised from making marginal safety improvements to AVs to avoid liability.[35] The law will therefore need to strike a balance in attributing responsibility. This balance will likely be achieved by apportioning liability based on ‘whether a human driver or the AV system was mainly operating the vehicle at the time of loss.’[36]
B Liability Risks
In the event of a collision, aggrieved parties may seek redress under the common law and Australian Consumer Law (ACL), thus exposing AV manufacturers, owners and occupants to significant liability risks.
1 Negligence Claim against Occupant or Owner
By virtue of the Motor Accident Insurance Act 1994 (Qld), Queensland operates a common law fault-based compulsory third party (CTP) insurance scheme. Under this scheme, an injured third party can claim for damages if they can establish negligence against an owner or driver of a motor vehicle. Liability risks may therefore arise for CTP Insurers of AV owners and occupants, subject to any full or partial contributions from parties such as the vehicle manufacturer.[37]
One possible cause of action is a negligence claim against the CTP Insurer of an AV occupant who is found in breach of a duty to pay attention.[38] For only partially automated AVs, one would reasonably anticipate that legal responsibility would reside with a person sitting in the driver’s seat who is able to resume control or is prompted to do so. However, it is likely that liability risks resulting from human inattention will devolve to manufacturers as technology shifts towards full AV autonomy.[39]
If an injury arose out of negligent AV maintenance, a claim may also be available against the vehicle owner or CTP Insurer.[40] This would be based on “a failure to service, maintain or upgrade the vehicle or its software in line with manufacturer instructions.”[41]
2 Claims against Manufacturers and Suppliers
Researchers have surmised that widespread usage of AVs will lead to increased manufacturer and retailer liability.[42] This is a strong possibility as plaintiffs in AV collisions could pursue a negligence claim, or seek an ACL remedy for defective products,[43] breach of consumer guarantees[44] or misleading or deceptive conduct.[45]
In relation to negligence, a manufacturer owes a duty to take reasonable care to avoid injury being suffered by AV users and reasonably foreseeable bystanders,[46] including road users and passengers.[47] To determine whether a manufacturer has exercised reasonable care, the Court will consider the state of technical and scientific knowledge available at the time of manufacture or distribution.[48]
Similarly, the ACL confers statutory rights of action on persons who suffer injury or loss caused by a manufacturer’s defective goods.[49] In defining ‘defect,’ the ACL adopts an objective test[50] based on consumer expectations.[51] As with negligence at common law, the safety defect provisions afford an exception based on the ‘state of scientific or technical knowledge’ at the time of supply of the goods.[52]
3 Uncertainty
Perhaps the greatest legal risk to potential legal actors is that collision liability is markedly uncertain, due to the amorphous legal tests embedded in the law and the complexities associated with attributing liability. Most problematically, standards at common law and under the ACL are unsuited to an AV context.[53] The ACL consumer expectation test has been criticised as being inapplicable to complex products that are not very familiar to consumers.[54] Similarly, application of the ‘state of scientific or technical knowledge’ test to AVs is riddled with uncertainty in a field experiencing rapid technological progress. Professor Bryant Smith notes that thinking about product defects ‘in terms of the decisions that [a vehicle] makes’ is anomalous.[55] Furthermore, there is uncertainty about whether to compare the driving standard of the AV to that of a human driver or another AV.[56] This uncertainty in the law might be compounded by complexities in attributing liability where there are joint liability or contributory negligence issues,[57] such as where a vehicle manufacturer claims against designers of AV software.[58]
OTHER RISKS
A Privacy
There is a strong impetus for insurers and manufacturers to consider potential privacy issues pertaining to AVs, due to possible risks of hacking, vehicle modification and unauthorised use of and access to data.[59] Most relevantly, the Privacy Act provides that where a manufacturer or insurer collects personal information about an individual for a particular purpose, they will be prohibited from using or disclosing that information for a secondary purpose.[60] This principle is subject to exceptions such as consent.[61] Relevant personal information for the purposes of the Privacy Act could include driver information, location data, or systems operation data.[62]
The use of event recordings data created by AVs is a key issue for manufacturers and insurers, as it could serve to clarify liability issues or improve an AV model’s future performance.[63] To diminish the risk of breaching the Privacy Act, manufacturers and insurers could obtain broad consents for use of an AV owner’s data,[64] and should limit use of information for a secondary purpose where possible.[65]
B Road Rules[66]
A final issue to consider is the uncertain liability risks for AV occupants under Queensland’s Road Rules.[67] Under the Road Rules, ‘a driver must not drive a vehicle unless the driver has proper control of the vehicle.’[68] It is also an offence for a person to drive without due care and attention.[69]
‘Driving’ has been defined in the case law as ‘causing a vehicle to move … down a road, and controlling the handlebars and brakes.’[70] The courts have held that pushing a vehicle with one hand on the steering wheel is not considered to be driving.[71] This presumably means the more passive act of occupying an AV would be insufficient, however further clarification is needed. Additionally, guidance is required as to how definitions such as ‘proper control’ and ‘vehicle’ will be applied in the context of AVs. In most States and Territories, police interpret ‘proper control’ to mean ‘having one hand on the steering wheel,’[72] but it is questionable how the term will be applied to AV usage.
SUGGESTED RECOMMENDATIONS
It
is recommended that Queensland adopt a no-fault insurance system, in order to alleviate
uncertainties and complexities associated with collision incidents, such as attribution
of liability issues.[73]
This reform would also safeguard individual drivers, stimulate marketplace
innovation and facilitate quicker provision of compensation to victims.[74]
Another recommendation is to clarify AV data ownership and the rights of manufacturers,
insurers and law enforcement to access event-recording data.[75]
Finally, traffic liability should be clarified for AV operators, and terms such
as ‘proper control’ and ‘driving’ in the Queensland Road Rules should be
amended to provide greater certainty in the AV context.[76]
[1] Selmer Bringsjord and Bettina Schimankski, ‘What is Artificial Intelligence? Psychometric AI as an Answer’ (2003) 18 Eighteenth International Joint Conference on Artificial Intelligence 887, 892; Michael Negnevitsky, Artificial Intelligence: A Guide to Intelligent Systems (Pearson Education Limited, 2nd ed, 2005) 18.
[2] Bartosz Brożek, Jaap Hage and Bipin Indurkhya, ‘Introduction to the special issue on machine law’ (2017) 25(3) Artificial Intelligence and Law 251, 252.
[3] Tom Standage, ‘How Does a Self-driving Car Work’, The Economist (online), 12 May 2015 <https://www.economist.com/blogs/economist-explains/2013/04/economist-explains-how-self-driving-car-works-driverless> visited 27 March 2018; Bill Robertson, ‘How Do Self-Driving Cars Work?’ (2017) 54(9) Science and Children 72, 73.
[4] Ashish Sukhadeve, How Machine Learning Will Drive Autonomous Vehicles (26 May 2017) Datafloq <https://datafloq.com/read/machine-learning-drive-autonomous-vehicles/3152> visited 27 March 2018.
[5] Owen Hayford et al, Driving into the Future: regulating driverless vehicles in Australia (19 September 2016) Clayton Utz <https://www.claytonutz.com/articledocuments/178/Clayton-Utz-Driving-into-the-future-regulating-driverless-vehicles-2016.pdf.aspx?Embed=Y> 7 visited 25 March 2018.
[6] Lisa Collingwood, ‘Privacy Implications and Liability Issues of Autonomous Vehicles’ (2017) 26(1) Information and Communications Technology Law 32, 32.
[7] Umar Zakir Abdul Hamid et al, ‘Current Collision Mitigation Technologies for Advanced Driver Assistance Systems’ (2016) 6(2) Perintis eJournal 78, 78.
[8] Ravi Shanker et al, ‘Autonomous Cars: Self-Driving the New Auto Industry Paradigm’(Blue Paper, Morgan Stanley, 6 November 2013) <https://orfe.princeton.edu/~alaink/SmartDrivingCars/PDFs/Nov2013MORGAN-STANLEY-BLUE-PAPER-AUTONOMOUS-CARS%EF%BC%9A-SELF-DRIVING-THE-NEW-AUTO-INDUSTRY-PARADIGM.pdf> visited 22 March 2018; see also: Hamid et al, above n 7, 78.
[9] David K Gibson, ‘Can we banish the phantom traffic jam?’, BBC (online), 28 April 2016 <http://www.bbc.com/autos/story/20160428-how-ai-will-solve-traffic-part-one> visited 27 March 2018; Shanker et al, above n 8.
[10] John Miller, Self-Driving Car Technology’s Benefits, Potential Risks and Solutions (20 August 2014) The Energy Collective <http://www.theenergycollective.com/jemiller_ep/464721/self-driving-car-technology-s-benefits-potential-risks-and-solutions> visited 22 March 2018.
[11] Patcharinee Tientrakool, Ya-Chi Ho and Nicholas Maxemchuk, ‘Highway Capacity Benefits from Using Vehicle-to-Vehicle Communication and Sensors for Collision Avoidance’ (Paper presented at Vehicular Technology Conference, San Francisco, 5 September 2011).
[12] Australian Driverless Vehicle Initiative, Submission to House of Representatives Standing Committee on Industry, Innovation, Science and Resources, Parliament of Australia, Inquiry into Social Issues relating to land-based driverless vehicles in Australia, February 2017, 9.
[13] Shanker et al, above n 8.
[14] Jon Fingas, ‘GM faces lawsuit over self-driving car collision’, Engadget (online), 28 January 2018 <https://www.engadget.com/2018/01/28/gm-faces-lawsuit-over-self-driving-car-collision/> visited 22 March 2018.
[15] Greg Miskelly, ‘Uber suspends self-driving car tests after vehicle hits and kills woman crossing the street in Arizona’, ABC News (online), 20 March 2018 <http://www.abc.net.au/news/2018-03-20/uber-suspends-self-driving-car-tests-after-fatal-crash/9565586> visited 22 March 2018; see also: Sam Levin, ‘Video released of Uber self-driving crash that killed woman in Arizona’, The Guardian (online), 22 March 2018 <https://www.theguardian.com/technology/2018/mar/22/video-released-of-uber-self-driving-crash-that-killed-woman-in-arizona> visited 22 March 2018.
[16] Carolyn Said, ‘Exclusive: Tempe police chief says early probe shows no fault by Uber’, San Francisco Chronicle (online), 19 March 2018 <https://www.sfchronicle.com/business/article/Exclusive-Tempe-police-chief-says-early-probe-12765481.php> visited 22 March 2018.
[17] Ethan Baron, ‘Blame Game: Self-Driving Car Crash highlights tricky legal question’, The Mercury News (online), January 23 2018 <https://www.mercurynews.com/2018/01/23/motorcyclist-hit-by-self-driving-car-in-s-f-sues-general-motors/> visited 22 March 2018.
[18] See, eg, Australian Competition and Consumer Commission, New Car Retailing Industry – a market study by the ACCC (14 December 2017) Australian Competition and Consumer Commission <https://www.accc.gov.au/system/files/New%20car%20retailing%20industry%20final%20report_0.pdf> visited 26 March 2018; see also: Australian Competition and Consumer Commission, Compliance and Enforcement Policy (20 February 2018) Australian Competition and Consumer Commission <https://www.accc.gov.au/system/files/D18-20423%20Enf%20-%20Admin%20Other%20-%20CLEAN%20VERSION%20final%20draft%20Combined%20Complia…%20%5Bfinal.%5D.pdf> visited 26 March 2018.
[19] Product Safety Australia, Browse all Transport Recalls (20 March 2018) Product Safety Australia <https://www.productsafety.gov.au/recalls/browse-all-recalls?f%5B0%5D=field_accc_psa_product_category%3A4792> visited 20 March 2018.
[20] The Hon Michael Sukkar MP, Explanatory Statement (28 February 2018) Product Safety Australia <https://www.productsafety.gov.au/system/files/Attachment%20B%20-%20Explanatory%20Statement%20Takata.pdf> visited 20 March 2018.
[21] See, eg, Chris Nichols, ‘Liability Could be Roadblock for Driverless Cars’, San Diego Union Tribune (online), 30 October 2013 <http://www.sandiegouniontribune.com/news/sdut-liability-driverless-car-transovation-google-2013oct30-story.html> visited 23 March 2018.
[22] See, eg, Brożek, Hage and Indurkhya, above n 2, 252.
[23] Jaap Hage, ‘Theoretical foundations for the responsibility of autonomous agents’ (2017) 25(3) Artificial Intelligence and Law 255, 271.
[24] Said, above n 16.
[25] Bartosz Brożek and Marek Jakubiec, ‘On the legal responsibility of autonomous machines’ (2017) 25(3) Artificial Intelligence and Law 293, 304.
[26] Frodo Podschwadek, ‘Do androids dream of normative endorsement?’ (2017) 25(3) Artificial Intelligence and Law 325, 339.
[27]Alexander Hevelke and Julian Nida-Rumelin, ‘Responsibility for Crashes of Autonomous Vehicles: An Ethical Analysis’ (2015) 21(3) Science and Engineering Ethics 619, 620.
[28] Maurice Schellekens, ‘Self-driving cars and the chilling effect of liability law’ (2015) 31 Computer Law and Security Review 506, 509.
[29] Volvo, ‘US urged to establish nationwide Federal guidelines for autonomous driving’ (Media Release, 7 October 2015) 1 <https://www.media.volvocars.com/global/en-gb/media/pressreleases/167975/us-urged-to-establish-nationwide-federal-guidelines-for-autonomous-driving> visited 28 March 2018.
[30] Gary Marchant and Rachel Lindor, ‘The Coming Collision Between Autonomous Vehicles and the Liability System’ (2012) 52 Santa Clara Law Review 1321, 1335-1336.
[31] Hevelke and Nida-Rumelin, above n 27, 619.
[32] Jeffrey Gurney, Sue My Car Not Me: Products Liability and Accidents Involving Autonomous Vehicles (JD Thesis, University of South Carolina, 2014) 247 <http://illinoisjltp.com/journal/wp-content/uploads/2013/12/Gurney.pdf> visited 28 March 2018.
[33] Sophia H. Duffy and Jamie Patrick Hopkins, ‘Sit, Stay, Drive: The Future of Autonomous Car Liability’ (2013) 16 Science and Technology Law Review 453, 453-455; Hevelke and Nida-Rumelin, above n 27, 626.
[34] Hevelke and Nida-Rumelin, above n 27, 619.
[35] Marchant and Lindor, above n 30, 1321.
[36] Munich RE, Autonomous Vehicles – Consideration for Personal and Commercial Lines Insurers (28 March 2016) Munich RE < https://www.munichre.com/site/mram-mobile/get/documents_E1725865033/mram/assetpool.mr_america/PDFs/3_Publications/Autonomous_Vehicles.pdf> visited 20 March 2018.
[37] Owen Hayford et al, above n 5, 17.
[38] Henry Silvester and Jacinta Daher, The unstoppable drive of automated vehicles – likely impacts on CTP Insurers (30 May 2017) Barry.Nilsson. Lawyers <https://www.bnlaw.com.au/page/Insights/Insurance_Alerts/Compulsory_Third_Party/The_Unstoppable_Drive_of_Automated_Vehicles_-_Likely_Impacts_on_CTP_Insurers/> visited 24 March 2018.
[39] Owen Hayford et al, above n 5, 2.
[40] Silvester and Daher, above n 38.
[41] Taylor Wessing, ‘Who’s in the driving seat? Driverless cars, liability and insurance’ (2017) Lexology 1, 1.
[42] James M. Anderson et al, Autonomous Vehicle Technology: A Guide for Policy Makers (RAND Corporation, 1st ed, 2016) xxii-xxiii.
[43] Competition and Consumer Act 2010 (Cth) sch 2 pt 3-5 div 1.
[44] Ibid pt 3-2, div 1. See particularly ss 54, 55, 60 and 61.
[45] Ibid pt 2-1.
[46] The Law Society of New South Wales, Submission to National Transport Commission, Regulatory barriers to more automated road and rail vehicles, 6 February 2017, 4.
[47] Owen Hayford et al, above n 5, 20.
[48] Ibid.
[49] Competition and Consumer Act 2010 (Cth) sch 2 ss 9, 138-139, 142(a)(c); The Law Society of New South Wales, above n 46, 4.
[50] Carey-Hazell v Getz Bros & Co Pty Ltd [2004] FCA 853 [186] (Kiefel J); Glendale Chemical Products Pty Ltd v ACCC (1999) 90 FCR 40, 47.
[51] Competition and Consumer Act 2010 (Cth) sch 2 s 9(1).
[52] Competition and Consumer Act 2010 (Cth) sch 2 s 142.
[53] Henry Prakken, ‘On the problem of making autonomous vehicles conform to traffic law’ (2017) 25(3) Artificial Intelligence and Law 341.
[54] Carolyn Sappideen and Prue Vines, Product Liability (October 28 2010) Thomson Reuters <https://legal.thomsonreuters.com.au/product/AU/files/720502336/chapter_23.pdf> visited 27 March 2018.
[55] Baron, above n 17.
[56] Schellekens, above n 28, 510.
[57] The Law Society of New South Wales, above n 46, 4; see also: Sappideen and Vines, above n 54.
[58] Silvester and Daher, above n 38.
[59] See generally National Transport Commission, ‘Regulatory reforms of automated vehicles’ (Policy Paper, National Transport Commission Australia, November 2016) <http://www.ntc.gov.au/Media/Reports/(32685218-7895-0E7C-ECF6-551177684E27).pdf> visited 25 March 2018; The Law Society of New South Wales, above n 46, 5; Christopher Dolan, ‘Self-Driving Cars and the Bumpy Road Ahead’ (2016) 1 American Association for Justice Trial Magazine 1; House of Representatives Standing Committee on Industry, Innovation, Science and Resources, Parliament of Australia, Social Issues Relating to Land-Based Automated Vehicles in Australia (2017) 30.
[60] Privacy Act 1988 (Cth) sch 1 APP 6.1; see also: Owen Hayford et al, above n 5, 29.
[61] Privacy Act 1988 (Cth) sch 1 APP 6.1-6.4.
[62] House of Representatives Standing Committee on Industry, Innovation, Science and Resources, above n 59, 17.
[63] Andrew Garza, “Look Ma, No Hands!” Wrinkles and Wrecks in the Age of Autonomous Vehicles (JD Thesis, University of Connecticut, 2009) 611-613 <http://www.boyleshaughnessy.com/Collateral/Documents/English-US/Garza%20No%20Hands.pdf> visited 28 March 2018; House of Representatives Standing Committee on Industry, Innovation, Science and Resources, above n 59, 16-17; Owen Hayford et al, above n 5, 29; National Transport Commission, above n 59, 63.
[64] Owen Hayford et al, above n 5, 29.
[65] Cara Bloom et al, ‘Self-driving cars and data collection: Privacy Perceptions of Networked Autonomous Vehicles’ (2017) 13 Proceedings of the Thirteenth Symposium on Usable Privacy and Security 357, 357; Senator Edward J. Markey, ‘Tracking and Hacking: Security and Privacy Gaps Put American Drives at Risk’ (Report, Markey Senate, 6 February 2015) 10.
[66] For a comprehensive overview of potential traffic law and criminal law liability in Queensland, see Kieran Tranter, ‘The Challenges of Autonomous Motor Vehicles for Queensland Road and Criminal Laws’ (2016) 16(2) QUT Law Revue 59.
[67] Transport Operations (Road Use Management—Road Rules) Regulation 2009 (Qld).
[68] Ibid s 297(1).
[69] Transport Operations (Road Use Management) Act 1995 (Qld) s 83.
[70] McNaughtan v Garland [1979] Qd R 240; Wallace v Major [1946] KB 473.
[72] National Transport Commission, above n 59, 33.
[73] Shanker et al, above n 8, 7.
[74] See, eg, Anderson et al, Autonomous Vehicle Technology: A Guide for Policy Makers (RAND Corporation, 1st ed, 2016).
[75] Owen Hayford et al, above n 5, 28; National Transport Commission, above n 59, 65.
[76] Tranter, above n 66, 81.
Australian Privacy Law Developments 