By Yeo Hwee Ying and Ruth Yeo
The advent of autonomous vehicles (‘AVs’) will disruptively transform the motor insurance sector. Traditional motor insurance in Singapore rests on the public policy of ensuring adequate coverage for all victims in road accidents. With AVs challenging the orthodox foundations of motor insurance, however, existing insurance doctrines will have to evolve to grapple with unfamiliar risks. We examine this in our article ‘Autonomous Vehicles and Insurance Law Principles: Navigating New Frontiers in Singapore’ published in the Singapore Academy of Law Journal.
The anticipated impact of driverless technology on the core insurance principles may be analysed as follows. First, the duties to disclose and not misrepresent facts will wane in relevance given the modern insurers’ ready access to material information via AVs equipped with telematics and other digital means. Furthermore, human risk factors will recede in importance while vehicular risk factors will become far more prominent. Insurers will hence find it difficult to rely on non-disclosure or misrepresentation to avoid contracts, as it will be problematic to assert that human factors are material or that they have induced the particular insurer to make the policy on the relevant terms. In contrast, information about vehicular risks can be readily obtained from manufacturers, rather than from the insured.
Second, the doctrines relating to terms, warranties and exclusions will take on heightened significance. Insurers commonly stipulate terms that mitigate the risks of untoward events. Regarding AVs, critical updates of the automation software as well as unauthorised tampering of the self-driving algorithms have been identified as being of critical concern. To manage risks, AV insurers may either: (a) specify exclusions for the purpose of avoiding or limiting liabilities when certain acts give rise to damages resulting from non-compliance by the insured; or (b) create prescriptive warranties. For example, undertakings to update automation software will likely be construed as warranties under the test in HIH Casualty & General Insurance v New Hampshire Insurance. Nevertheless, this software-update warranty can still be deemed a term delimiting risk suspending cover until the insured has remedied the breach by installing the requisite software updates.
Third, by reducing the driver’s role, the issues of causation and fault attribution will be further complicated. For instance, when a semi-AV (i.e. an AV which still requires a driver to be on stand-by to take over the wheel when prompted) crashes, complexity increases as it must be determined whether the incident is due to the carelessness of the human driver or some limitation of the automation software. Problems arise when either of the two causes has been explicitly excluded from coverage, as the exclusion clause will take priority and the insurer will consequently avoid liability for the insured’s loss. Nevertheless, the courts have generally construed exclusion clauses for negligence narrowly. Besides this, the onus of proving that the causative event falls within policy terms lies on the insured. However, a layman insured would find it difficult to adduce technical evidence of the semi-AV’s malfunction. As such, a suggested alternative is to opt for a rebuttable presumption where semi-AV collisions are prima facie attributable to software or hardware malfunctions. The burden of proving otherwise would thus fall on the semi-AV manufacturer, which would arguably incentivise manufacturers to improve safety enhancements for their semi-AV designs.
Fourth, regarding advanced AVs (which are fully autonomous), an alternative approach has been suggested. This approach is to use agency principles to construe an AV as an agent of the owner. However, the agency approach is not likely to attract traction, because imputing blame on the owner when the human element is manifestly absent for all vehicle-control operations appears to be unfair in principle. Apart from the agency approach, an alternative suggestion is to regard the autonomous entity as an instrument (or extension) of the owner. Indeed, common law cases involving digital machines employed on behalf of human contractors have been resolved on contract law principles without recourse to complex agency principles.
Fifth, the notification term, stipulating that the insured must inform the insurer as soon as possible in the event of any mishap, is likely to decline in importance because of the automated features incorporated into the vehicle design to facilitate incident reporting and claims submission.
Finally, the number of fraudulent claims is expected to fall significantly, as the wealth of vehicle-operation details stored in the AV’s on-board data recorder enables the insurance claims processor to double-check the information reported in the insured’s claim. Furthermore, perpetrating fraudulent claims will become more difficult as automation software cannot, unless hacked, be readily manipulated to cause a collision.
Regarding the present AV landscape locally, Singapore has chosen the approach that AV regulation should be enacted progressively in stages. Thus far, Parliament has only added interim provisions to the Road Traffic Act in 2017, meant exclusively to deal with the initial phases of AV testing. The interim provisions dictate compulsory cover for AVs. However, should the insurer invalidate the cover using the defences in its arsenal, a tort victim in an AV-involved accident is much less protected than their counterpart victims injured by a conventional motor vehicle, which is governed by the Motor Vehicles Act.
In future, however, a broader range of collision scenarios can be anticipated. Legislation must be passed to meet the forthcoming plans for full-fledged AV deployment. Critically, the government must decide the type of liability regime Singapore will choose to adopt. The Law Reform Committee of the Singapore Academy of Law has issued a report which concluded that although the optimal basis for an AV regulatory framework remains unclear, the preferred approach may be to retain the existing system with targeted and necessary modifications. This approach, if adopted, will mean that the system will continue to be undergirded by insurance law principles. The usual defences of the insurer will thus remain relevant. Insurance stakeholders will have to carefully consider the issues surrounding the interfacing of existing insurance principles with the full-scale commercialisation of AVs, to ensure that the insurance doctrines remain relevant in the face of this disruptive technology.
Keywords: Driverless Technology, Artificial Intelligence, Insurance Doctrines, Tort Victims
AUTHOR INFORMATION
Yeo Hwee Ying is an Associate Professor at NUS, Faculty of Law.
Email: lawyeohy@nus.edu.sg
Ruth Yeo is a Senior Lecturer at School of Law, Singapore University of Social Sciences.
Email: ruthyeosa@suss.edu.sg