Written evidence submitted by Clyde & Co LLP (SDV0050)


1                          Introduction

Inquiry: Self-driving vehicles

1.1                   The Committee is particularly interested in receiving written evidence that addresses:

(a)         likely uses, including private cars, public transport and commercial vehicles, and levels of automation;

(b)         progress of research and trials in the UK and abroad;

(c)          potential implications for infrastructure, both physical and digital;

(d)         the regulatory framework, including legal status, approval and authorisation processes and insurance;

(e)         safety and perceptions of safety, including the relationship with other road users such as pedestrians, cyclists and conventionally driven vehicles; 

(f)           the role of Government and other responsible bodies, such as National Highways and local authorities; and

(g)         potential effects on patterns of car ownership, vehicle taxation and decarbonisation in the car market.


2                          The liability regime for ‘self driving’ vehicles and the need to address uninsured use

2.1                   Part 1 of the Automated and Electric Vehicles Act 2018 (AEVA) deals only with liability of insurers when an automated vehicle (AV) is driving itself. It imposes strict liability on a motor insurer for accidents caused by a vehicle driving itself. That concept is defined as “operating in a mode in which it is not being controlled, and does not need to be monitored, by an individual”. This strict liability is wider than a motor insurer’s liability for conventional driving because that requires proof of negligence by a (human) driver whereas the AEVA regime requires merely showing that the vehicle was driving itself at the time.

2.2                   This approach in AEVA means that anyone who suffers loss or injury caused by an insured automated vehicle driving itself is able to make a claim against a motor insurer in more or less the usual way, rather than having to undertake a much more complex product liability claim against the vehicle manufacturer.  The Explanatory Notes (EN) to part 1 of AEVA summarise the position thus:

11. - Automated vehicles are those that have the capability of driving themselves without human oversight or intervention for some, or all, of a journey. In an automated vehicle the driver can, in at least some circumstances or situations, hand all control and responsibility to the vehicle and effectively become a passenger, using it in automated mode. United Kingdom law on compulsory motor insurance has focused historically on ensuring that victims of road traffic collisions are compensated quickly and fairly. In the case of an automated vehicle being operated in automated mode, however, accidents could take place not as a result of human fault, but because of a failure in the vehicle itself, for which the only recourse available to an otherwise uninsured victim might be to sue the manufacturer through the courts. This Part extends compulsory motor vehicle insurance to cover the use of automated vehicles in automated mode, so that victims (including the ‘driver’) of an accident caused by a fault in the automated vehicle itself will be covered by the compulsory insurance in place on the vehicle. The insurer would be initially liable to pay compensation to any victim, including to the driver who had legitimately handed control to the vehicle. The insurer then would have the right to recover costs from the liable party under existing common and product law.” 

2.3                   The EN also explain that this regime, at sections 2 (insurer’s strict liability) and 5 (insurer’s right of recovery) of the Act, presupposes that the vehicle is insured:

16. - Section 2(1) places first instance liability on the insurer for an accident caused by an automated vehicle on a road or other public place in Great Britain if the vehicle is insured.”

2.4                   Consequently, if an automated vehicle is uninsured - for whatever reason - the strict liability regime will not apply and the injured person’s rights to recover could be compromised. The current arrangements operated by the Motor Insurers’ Bureau in respect of uninsured drivers cannot apply because there is no (human) ‘driver’ involved.

2.5                   This gap in protection was recognised by the Law Commissions in their joint report of January 2022. Its recommendation 72 was that the government should “put in place measures to provide compensation in respect of uninsured authorised vehicles, equivalent to that provided for under section 2(1) of Automated and Electric Vehicles Act 2018.”

2.6                   We support this recommendation and would add that those measures should include introducing rights of recovery, equivalent to those in section 5 of the Act, in favour of any entity which might be tasked with compensating for the harmful consequences of uninsured automated driving. In our view, it would be fair and reasonable for such an entity to be able to seek contributions from other parties, such as manufacturers, in the same way as insurers would be able to do as a consequence of section 5.

3                          Data relating to automated operation

3.1                   It will be clear from the above that the liability regime set out in AEVA turns on establishing the operating mode of the vehicle at the time of the incident. For the regime of insurer strict liability and subsequent recovery (against, for example, a manufacturer) to work effectively and efficiently, we submit that a core minimum set of data relating to the vehicle’s operating mode should be stored, made available to insurers, and retained for an appropriate period (given the likelihood of delays in claims notifications).

3.2                   Significant work has been done in the insurance industry with other stakeholders to develop such a core minimum set of data. This has become known as the data storage system for automated driving or DSSAD.

3.3                   Requirements of automated driving systems to store, share and retain data (in DSSAD specification) should be part of any pre-market authorisation process. This again is a recommendation made by the Law Commissions (at paragraphs 5.96 & 13.40 – 13.62 of their final report) and would ensure that only those automated driving systems with adequate data sharing capabilities would be approved for road use.

3.4                   In our view, the underlying policy goals of the AEVA liability regime are:

(a)         to protect those harmed by automated driving by introducing strict liability on insurers (section 2) in the first instance; and

(b)         to apportion fairly the financial risks of harm caused by automated driving to those responsible for the operating systems involved via the mechanism of the secondary recovery by insurers (section 5) from the likes of manufacturers or software houses.

3.5                   We therefore support the Law Commissions’ recommendations 73 and 74 which respectively propose new legislation on automated driving to:

(a)         “[empower] the authorisation authority [to] require data to be collected and stored to process insurance claims”, and

(b)         “impose a duty on those controlling AV data to disclose data to insurers, where the data is necessary to decide claims fairly and accurately.”

3.6                   In our view it may however be necessary to put further measures in place relating to data disclosure. For example, these should ensure that the data shared for claims handling purposes is relevant and readily accessible (rather than being a system ‘dump’ of complex data or code from the vehicle) and that there are appropriate sanctions for manipulating or tampering with data before disclosure. It may also be necessary to address the possibility that the manufacturer ceases trading before claims are presented.

3.7                   We would therefore support the call from the insurance industry for the necessary data to be disclosed to and retained by a neutral central server.

4                          Security and systemic hacking

4.1                   We note that “cybersecurity and hacking” were expressly excluded from the Law Commission’s terms of reference for its work on automated driving on the basis that these (and other) “policy areas of AV reform are being led directly by the UK government”, paragraph 1.15 of the report.

4.2                   These are critical issues. Malicious systemic hacking of a fleet or of particular models of automated vehicles seem to us to be of an entirely different nature than the prospect of individuals tampering with software that is in part addressed in AEVA at section 4 We would therefore question whether the strict liability regime in AEVA is an appropriate response to the harmful consequences of malicious, possibly terrorist, systemic hacking of automated vehicles (or even, in the case of connected vehicles, systemic hacking of the connected road infrastructure).

4.3                   Alternative solutions may be required because of the very wide potential exposures. Those might involve government bodies or funds or even innovative reinsurance approaches modelled, for example, on the protection afforded by Pool Re for property terrorism risks.

5                          Software updating

5.1                   The regulatory framework for automated driving must address the need to update software in a clear and unambiguous manner. It should set out who should be subject to legal duties to update software and must also deal with the consequences of failure to update, something which is set out in part at section 4 of AEVA.

5.2                   We can foresee that insurers could make timely updating of software a condition of policy coverage and/or introduce a requirement to notify them that updates have been completed.   These may not be easy issues to deal with where software updates are provided ‘over the air’ and even without the policyholder’s knowledge.

5.3                   This area could potentially pose problems involving insurance contract law concepts of disclosure and fair presentation of risk, particularly because the capabilities of the insured AV - and the risks those represent - could change rapidly and frequently. These circumstances are materially different to those applying to conventional vehicles, in that:

(a)         their performance does not generally change; and

(b)         insurers will require notice of modifications in any event.

5.4                   An alternative idea to requiring policyholders to notify updates to insurers might be some form of central repository which receives information from manufacturers about updates and provides that information to insurers. That might be carried out by a body such as DVLA or the ‘neutral server’ mentioned above. We put this forward for discussion only as we are unaware of its feasibility in practice.

6                          The second hand market

6.1                   Although there are difficult issues raised in insuring automated vehicles and in dealing with claims, it seems to us that those are more readily solvable when thinking about new automated vehicles. Relatively clear terms setting out respective rights, duties and obligations on issues such as maintenance, data and software should be capable of being agreed with owners and policyholders when cars are first bought, leased or insured. It may even be that some of those terms reflect the regulatory regime for pre-market and in-use authorisation of automated driving systems.

6.2                   It seems to us, however, that there is likely to be less opportunity for such matters to be addressed when automated vehicles become available on the second hand market. It is not clear to us that solutions for this circumstance are being actively discussed, although we accept that policing the second hand AV market could fall within the remit of the ‘in use’ AV regulator recommended by the Law Commission. Nevertheless, we would recommend that in any new regulatory regime for automated driving there should be a clear and unequivocal statement that it covers the second hand market.