Guest contributor Daniel West (Head of Product Liability at Horwich Farrelly) considers the role product liability is likely to play in resolving claims involving collisions caused by autonomous vehicles (AVs) driving themselves.


When discussing civil liability, the Law Commission (LC) said in its joint report on AVs (published on 26.01.22; see Automated Vehicles | Law Commission) that “product liability law is likely to play only a limited role in the regulation of self-driving” (see 13.27). Some might think that a rather bold statement – particularly given the considerable uncertainty as to how the AV industry will develop and how both the AV and insurance industries will respond to compensation claims under the Automated and Electric Vehicles Act 2018 (the AEVA). Below, we have sought to unpick why the LC appears to think that product liability will have such a limited part to play, before considering the ways in which (we say) product liability may in fact have a more prominent role once we start to see automated driving systems in use on UK roads – particularly if current sales, insurance and ownership models are maintained.

The legislative framework

The LC considered civil liability in the context of AVs in chapter 13 of its report. It outlined the legislative regime introduced by the AEVA, including that Section 2(1) of the AEVA introduces a form of strict liability on insurers of AVs for any ‘damage’ resulting from a collision caused by an AV driving itself. ‘Damage’ is defined as death or personal injury or any property damage save for certain exceptions, which includes damage to the automated vehicle itself. The type of person who can make a claim under the AEVA is not restricted and includes the ‘user-in-charge’, i.e., the person in the vehicle who can operate the controls and who is required to take over following a transition demand.

Section 5(1) of the AEVA then permits an insurer which has settled a claim under Section 2(1) (whether by judgment or agreement) to claim against “any other person liable to the injured party in respect of the accident”. (It’s worth noting here that the insured would have had the same right of recovery under the Civil Liability (Contribution) Act 1978, but the AEVA takes away this right in lieu of the right under Section 5(1).)

The Law Commission’s view on the role of product liability

The LC went on to consider the possibility insurers will bring claims against AV manufacturers under the Consumer Protection Act 1987 (the CPA). It is in this context the LC reached the view that product liability will only play a limited role here. Its reasoning appears to have been as follows:

1. The LC noted (at 13.6) that both the driver and AV’s liability will be provided for under the same policy (per the amended Section 145 of the Road Traffic Act 1988) which it said will prevent disputes about whether the driver or AV was in control at the time of the incident. The implication here appears to be that insurers will pay out claims to injured parties regardless of whether a collision was caused by the AV driving itself, since this should only impact whether the insurer can make a recovery.

2. The LC considered that “Individuals are unlikely to litigate, given the immense difficulty and cost of bringing product liability claims, and their poor success rate” (see 13.27).

3. With respect to claims brought by insurers against AV manufacturers, the LC has said that the parties will often “be able to resolve matters without recourse to the law” (see 13.28). For example, the LC suggested that if a regulator has already found that an AV is in breach of the relevant regulations, then this would be highly influential in any civil proceedings.

We have considered each of these assertions in turn below.

What role might product liability play?

Is it likely insurers will pay out to injured parties regardless of whether the driver or AV was controlling the vehicle?

In determining how an insurer might respond to a claim arising from a collision where there is uncertainty as to whether the driver or the AV was in control, it is useful to have in mind the types of damage that might be sustained when an AV is involved in a collision and to consider how a policy would respond to each type of damage.

In the table below, we have (for completeness) considered how both a fully comprehensive policy and a third-party policy would respond to each type of damage; but we would flag here that there may not in fact be much of a market for purchasing third party policies in connection with AVs. There is nothing in the Road Traffic Act 1988 prohibiting third party policies for AVs but given the likely value of an AV, it seems likely most (if not all) users will purchase fully comprehensive policies.

We can see from the above table that the question of whether the driver or the AV was in control will make little difference to the insurer when dealing with claims by third parties (it will pay either way) or for damage to the AV itself (it will pay on fully comprehensive policy either way but not on a third-party policy either way). To this extent, the LC is perhaps right in its assertion that insurers will be inclined to pay out such claims regardless of whether the driver or the AV was in control.

However, when dealing with injuries sustained by the driver or user-in-charge, there will be a key difference which is that any payment by the insurer would be entirely dependent on whether the driver or AV was in control. As a result, at least in some cases the insurer may delay paying such a claim until it knows whether the driver or AV was in control; because, to do otherwise, the insurer risks paying for a claim on the assumption the AV was to blame only to learn later that the AV manufacturer can show otherwise – in which case the insurer would have paid when it was not obliged to do so, but with no way to recover its outlay (other than by seeking to reclaim the payment from the driver/user-charge).

It’s also worth noting here that some fully comprehensive polices include an extension for the policyholder to drive another car not owned by them (‘Driving Other Cars’ or ‘DOC’ cover). In these circumstances, the extension provides third party cover only. But it’s not clear how such an extension would apply in the context of an AV owner using someone else’s AV whilst driving itself, not least because the wordings of most DOC extensions refer to the policyholder “driving” another vehicle. (It may be insurers will adapt their policy wordings or will not be prepared to offer DOC extensions on AV policies.) If a policyholder with DOC cover were to use someone else’s AV and was involved in a collision, then (again) the question of whether the driver or the AV was in control will be important to the insurer because this will determine whether the DOC insurer or the AV insurer is liable. Again, the AV insurer may delay paying the claim until it knows the AV manufacturer’s position.

Both examples above could result in the type of protracted, multi-party claim often seen in product liability litigation, i.e., precisely what the government had hoped to avoid in enacting the AEVA.

Are individuals likely to bring product liability claims?

As above, the LC says individuals are unlikely to bring product liability claims, essentially because such claims are difficult and have a poor success rate. However, there appears to be a more obvious reason, which is that an individual will already have a straightforward claim against the insurer of the AV under the AEVA. (This was of course one of the AEVA’s key objectives.)

However, that is not to say there is no scope for individuals to bring product liability claims. For example, a policyholder who has been refused indemnity or who only has the benefit of third-party cover may have to consider a breach of contract claim against the dealership which supplied him the AV for the cost of repairing or replacing it (particularly bearing in mind the cost of an AV).

And where there is merit in doing so, it is hard to imagine that claimants will be put off by any apparent perception that product liability claims are difficult. Certainly, claimants do not appear to be so dissuaded in other areas of product liability litigation.

Are insurers and AV manufacturers likely to resolve matters without recourse to the law?

The LC’s comments that commercial entities will be able to resolve disputes amicably belies the experience of product liability practitioners in other industries. In fact, that experience suggests litigation may well be inevitable, sometimes protracted and could involve the injured parties (contrary to the best intentions of the AEVA). We say this for the following reasons:

1. The first point to make is that there will likely be many occasions when a collision occurs when an AV is driving itself but where it is less clear that the collision was caused by the AV driving itself. For instance, collisions may be caused (at least in part) by the driver of another vehicle (or the actions of another AV), the actions of a pedestrian or the action or inaction of the user-in-charge (or a passenger). As in motor claims generally, there is clearly scope for much dispute between insurers and recovery targets as to the cause of a collision – and it is perhaps naïve to suggest this will not drive litigation.

2. The LC appears to have equated product liability law with claims under the CPA against manufacturers (see 13.24). But in fact, the product liability regime in the UK is made up of multiple causes of action including negligence and breach of contract (as well as claims under the CPA) and will often involve claims against multiple parties in the supply chain. This is important because in the event an insurer has settled a claim arising from a collision caused by an AV driving itself, it can sue any other party liable to the injured parties and will often have several different recovery options under Section 5(1). The insurer will no doubt make an assessment in each case as to who it can sue based on what it will be required to prove and any relevant limitation periods.

3. Even within the strict liability regime provided by the CPA, the Act offers up several possible defendants including as follows:

  • The manufacturer (producer) of the AV and/or software.
  • Any party which imported the AV and/or software into the UK.
  • Any party which has held itself out as the producer of the AV and/or software by using its name or logo in connection with either product.
  • Any party which supplied the AV and/or software (but only if this party is asked to identify one of the other parties listed above and it fails to do so within a reasonable period).

We can see from the above list that if the AV had been imported into the UK, then it is almost inevitable that the insurer would choose to seek a recovery from the UK importer (rather than the foreign AV manufacturer) in order to avoid grappling with jurisdictional issues, etc. Then, once it makes such a claim, the UK importer will likely look to make its own claim against another party (either the manufacturer or next party in the supply chain). In these circumstances, it is not difficult to see the potential for multi-party, cross border product liability litigation to arise.

4. For losses sustained by the AV owner (such as injury to them and damage to the AV itself), the insurer may also be able to consider a claim against the party which supplied the AV and/or software based on their potential liability for breach of contract – having potentially supplied a product in breach of implied terms that it would be of satisfactory quality and fit for purpose. The supplier may be the dealership, a finance company or a previous owner. In some circumstances, particularly where the owner is the only person to have sustained any losses, this contractual claim may in fact be the preferred option. That is because in all cases the insurer will not be able to bring a claim under Section 5(1) of the AEVA for damage to the AV itself, because (as above) Section 2(3) excludes this from the definition of ‘damage’ (and in any event because any other party would not be liable to the injured party under the CPA or in negligence for damage to the AV itself, since this would be pure economic loss). This means that in all claims, the insurer will need to recover any amount it pays in respect of damage to the AV itself under alternative means including, for example, a subrogated claim for breach of contract against the dealership. If the insurer must sue the dealership anyway, then it may be easier to seek all its losses from the dealership rather than pursue two separate parties.

5. In fact, in some circumstances, a claim based in contract may be the only option. That is because whilst the AEVA and Limitation Act 1980 provide the insurer with 2 years from the date of judgment or settlement to bring a claim against another party under Section 5(1), it may not be able to do so based on that party’s liability under the CPA if a period of over 10 years has elapsed since that party first supplied the AV to another. That is because there is a 10 year ‘long-stop’ for CPA claims which operates to extinguish the right of action entirely following the expiry of 10 years from the date the product was originally supplied. Interestingly, the Civil Liability (Contribution) Act 1978 explicitly confirms at Section 1(3) that a party cannot bring a contribution claim if a limitation period has extinguished the right on which the claim is based. It’s not yet clear if the same will apply to a claim under Section 5(1) of the AEVA, but there is no reason to think it would not, and if it does apply then insurers will have to consider alternative options for collisions involving AVs over 10 years old. (It’s currently estimated that electric cars can last between 10 and 20 years but we do not know yet what the lifespan of a typical AV will be.)

6. Like the position with UK importers (as outlined above), a dealership facing a claim based on its liability in contract will likely also look to make its own recovery claim – so, again, multi-party litigation is a real risk.

7. It is common for commercial parties in multi-party litigation to have agreed (or sought to agree) terms and conditions and this can have a significant impact on any subsequent civil claim. For example, there may be clauses which seek to limit or exclude a party’s liability to another, or which provide that any dispute be dealt with in a particular jurisdiction and in accordance with a particular governing law. Again, this will only serve to drive litigation between the commercial parties.

Other product liability issues

There is a range of other issues likely to be raised in product liability claims arising from collisions involving AVs, which may well promote litigation, as we have outlined below:

1. The ‘development risks defence’ under Section 4(1)(e) of the CPA provides that it will be a defence for an AV manufacturer to show “that the state of scientific and technical knowledge at the relevant time was not such that a producer of products of the same description as the product in question might be expected to have discovered the defect if it had existed in his products while they were under his control”. This defence is intended to protect a manufacturer in circumstances where it could not have known (based on the scientific and technical knowledge available at the time) that a defect existed in its product. It’s certainly possible AV manufacturers will seek to rely on this defence in the context of AVs, particularly given that the scientific and technical knowledge available is changing continuously.

2. Similarly, the courts have also made clear in recent product liability cases (see Wilkes v DePuy (2016) EWHC 3096 (QB) and Gee and others v DePuy International Limited [2018] EWHC 1208 (QB)) that when considering whether there is a defect in a product, they will consider factors such as the product’s costs, its risk-benefit profile and the ease and extent to which a safety risk can be eliminated. The courts have confirmed that just because a safer design can be envisaged this does not mean a product is defective and that they are prepared to weigh the potential benefits of a product against the risks it might cause injury. Both those cases concerned metal-on-metal hip replacements, but there’s no reason the same analysis would not be applied to automated driving systems, with AV manufacturers potentially arguing that a system should be considered safe, despite having caused some collisions, because it causes less collisions than a traditional vehicle.

3. There is uncertainty at present as to how traditional product liability laws will be applied to modern, connected products, like AVs, which can change over time, including because of software updates (or even through AI machine learning). Traditionally, the safety of a product is assessed at the point it’s put into circulation, but clearly that is not satisfactory where AV manufacturers make significant changes to a product after it’s supplied, via frequent (sometimes safety critical) software updates. Such a software update could introduce a flaw and, in these circumstances, it’s not clear how limitation will apply, particularly for an update which is issued after 6 years (the deadline for a breach of contract claim) or 10 years (the deadline for a claim under the CPA). It’s also not clear for how long AV manufacturers will support their AVs, whether there will be any obligation on them to do so and whether there will be cut-offs for older models (as seen with, e.g., iPhones). These types of issue are under review by the UK’s Office for Product Safety & Standards and the EU with the possibility that reforms to product liability laws will follow.

Conclusion / commentary

It’s too early to say how prominent a role product liability will play in claims arising from the use of AVs, but we hope the above demonstrates there is at least potential for significant product liability litigation to arise.

Indeed, it may be the possibility of such claims will drive some motor manufacturers towards different sales models involving, for example, direct sales by manufacturers to consumers (a model which Tesla already follows and which the industry is moving towards) and/or manufacturers insuring their own vehicles (something Tesla has already started to do in the US). Such sales models (coupled with a potential move away from private ownership towards, e.g., subscription models) may help to minimise the incidence of complex, multi-party product liability claims and would allow AV manufacturers to control any litigation.

At this stage, it’s only possible to speculate how the AV and insurance industries and, in turn, the legal landscape will develop in the coming years.

Daniel West, Head of Product Liability, Horwich Farrelly

You may also like...