Insurance at the front line of automated vehicle regulation

A familiar starting point – but a fundamentally different risk model

As with conventional vehicles, all AVs will be required to carry compulsory motor insurance. The continuity is deliberate. However, the underlying liability model departs significantly from traditional driver-centric frameworks.

For vehicles with user-in-charge (UIC) features, insurance must cover both human driving (when the automated system is disengaged) and automated driving (when it is engaged). Responsibility for ensuring cover remains with the registered keeper. For no-user-in-charge (NUIC) vehicles, the position is more nuanced. Responsibility for insurance will depend on vehicle ownership and the conditions of the NUIC operator's licence. A vehicle is likely to have one UIC or one NUIC feature for a particular function, although both may exist across different use cases.

This deliberate flexibility enables a range of commercial models, but it also introduces complexity. Insurance allocation becomes a licensing and contractual issue, rather than one dictated exhaustively by statute.

Insurer liability under AEVA: victim-first, recovery later

The framework preserves the Automated and Electric Vehicles Act 2018 (AEVA) model. Insurers compensate victims first, with responsibility resolved via recovery actions against other liable parties, including other at-fault drivers and vehicle producers under product liability principles and the Consumer Protection Act 1987. Determining causation in practice, however, may be more complex in AV cases, particularly where responsibility may need to be apportioned between system behaviour, human intervention, and external road conditions.

Contributory negligence is likely to remain relevant, including failures such as not installing software updates, unsafe loading or conduct of other road users. Its precise application will develop case-by-case by the Courts.

Data replaces the driver as the evidential cornerstone

Perhaps the most significant shift for insurers is evidential. Traditional claims models rely heavily on driver accounts and witness testimony. Consistent with emerging discussions under the United Nations Economic Commission for Europe (UNECE) and data concepts such as Data Storage Systems for Automated Driving (DSSAD) automated vehicles may be expected to record and store data in ways that support incident analysis and claims assessment. This includes timestamps for system activation and deactivation, transition demands, collision detection, and more detailed sensor and vehicle dynamics data for safety-critical events.

Insurers will see this data as essential. Liability turns on whether the automated driving system was active, how it behaved, and whether any human intervention occurred. In practice, insurers are likely to focus in particular on whether the ADS was engaged, whether a transition demand was issued, and how the vehicle responded. The framework and consultation materials signal the importance of timely and standardised access to relevant data, but the precise mechanism and scope of that access remain under consideration. There is also an unresolved debate as to whether insurers should receive only limited status data or broader operational datasets.

Mandatory disclosure to insurers – a pivotal unresolved issue

The Law Commissions recommended a statutory duty on those controlling AV data to disclose information to insurers where necessary to determine claims. While the AV Act stops short of imposing that duty directly, it enables authorisation requirements and operator licensing conditions to mandate data sharing with insurers. Government has accepted the recommendation in principle and is considering whether, and how to introduce formal disclosure obligations. This is a critical inflection point and will determine whether insurers need to rely on contractual consent through policy terms.

Retention periods shaped by insurance litigation

Data retention is similarly informed by insurance realities. The Law Commissions proposed retention for 39 months, reflecting the three-year statute of limitations period for bringing a legal claim for personal injury, plus an additional three-month buffer to enable data to be retrieved and processed to respond to the claim. Government is considering mandating this domestically, while acknowledging storage, security, and data protection trade-offs.

The proposed alignment of technical data policy with litigation timelines would underscore how deeply insurance considerations may shape regulatory design.

Insurance as a regulatory assurance tool

Beyond motor insurance, also raises broader questions about how insurance may interact with regulatory concepts such as financial standing, operational responsibility and market readiness. In that sense, insurance may function not only as a mechanism for risk transfer, but also as one indicator of whether an operator or authorised entity can support safe and reliable deployment. The framework does not currently prescribe additional insurance requirements of this kind, however, and the extent of any broader insurance expectations remains to be worked through.

Looking ahead

The regulatory direction is clear. Insurance is no longer ancillary to automated mobility; it is a critical consideration in determining the broader UK regulatory framework. Insurers will sit at the front line of victim compensation, data interpretation, and risk pricing, while manufacturers and operators will need to navigate exposure through recovery actions and regulatory scrutiny.

For legal and insurance practitioners, the key challenges lie in data governance, allocation of responsibility across complex operating models, and the interaction between regulatory requirements and private insurance contracts. How these issues are resolved will shape not only claims outcomes, but the pace and confidence with which automated vehicles are deployed on UK roads. Insurers will also need to maintain realistic expectations about the scope, accessibility and cost of DSSAD data.

Authored by Karishma Paroha.