Autonomous vehicles will stretch the limits of liability

The way in which people and goods travel is changing rapidly. The shift towards transport in which machines are increasingly taking control from humans raises fundamental considerations around public safety, ethical dilemmas and where liability rests when accidents occur. It also means a major shift in the amount of data that is collected by vehicles, and how that data is stored and used.

Gearing up for this new era of transportation presents many challenges. The application of autonomous vehicle technology across different transport sectors will present unique obstacles and opportunities. The focus of this article is on some of those faced by the motor insurance market and the cargo and freight markets.

The motor insurance market

Widespread adoption of autonomous vehicles will herald dramatic changes for the motor insurance market. Insurers will be faced with strategic challenges in continuing to support the classic insurance model towards new risk models in which the liability moves towards product manufacturers. Removing the scope for human error behind the wheel, as seen with full automation, removes the driver from the liability equation and weakens the relationship between the driver and the insurer. In its place, the market will have to move more towards a product liability focus.

Insurers will need to develop autonomous vehicle insurance propositions as they see changes in consumer appetite for insurance products. We are likely to witness a move away from annual renewals towards on-the-go insurance and the growth of more transactional relationships where less is understood about the risk profiles of individual drivers.

It is anticipated that the new generation of road vehicles will encounter less wear and tear. In theory, this should reduce both maintenance costs as well as further lowering insurance pay-outs and premiums. However, maintenance and repair costs may still be high as the sophistication of the technology increases with newer models of vehicles, and because usually only the manufacturer can conduct the required servicing and update the software. This, in turn, means insurance premiums can actually increase significantly (certainly in the short term).

Connected and autonomous vehicles will create a massive increase in data, which may help to improve risk pricing and ultimately reduce premiums, but only if such data is not siloed.   Access to data is vital to prevent occlusion of information, and to improve underwriting and the provision of insurance for autonomous vehicles.

Data privacy and hacking are understandably, significant concerns for consumers. To ensure that insurers are central to providing risk management solutions to their customers, they will need to take an active lead in autonomous vehicle initiatives – as is already being seen with those investing in mobility research and development.  

Fleets and insurers are only just getting to grips with the collection of data from the latest generation of vehicles post-accident. Progressing to the level of technology and data that will be generated by level 4 or 5 vehicles will be a massive change that will need a significant level of resourcing which in the short term is likely to keep insurance premiums higher until the use of such vehicles becomes the norm rather than the exception.

At the moment vehicle data (as opposed to drivers own in-vehicle video record) is rarely obtained in lower value accidents, perhaps with accidents involving fleet or commercial vehicles as an exception, where electronic tachographs may be reviewed. Only in more serious road accidents will the police typically undertake a fuller investigation that may include accessing the vehicle data. This may then trigger vehicle owners / insurers to look to access the data themselves. A move towards stage 4/5 vehicles will need to see a change in approach by claims teams as the data capture, assessment (and indeed storage) of the data during the life cycle of a claim will become the norm and play an even greater part in the assessment of liability, in particular.

Cargo and freight markets

The cargo and freight markets have embraced rapidly developing autonomous vehicle technology.  It presents the opportunity to reduce substantially the costs involved in providing supply chain solutions and removes some of the risks inherent in human intervention. However, it also introduces some of its own risks and problems.

A number of international conventions governs the international carriage of goods. The conventions which govern the carriage of goods by air (Warsaw and Montreal) were drafted at the beginning of the 20th century; the CMR - which governs road carriage - is based on rail freight conventions which predate the CMR’s introduction in 1956 (collectively, the Conventions).

The Conventions all state that the carrier is responsible for the acts and omissions of those persons of whose services they make use in the performance of the carriage. Insofar as this applies to drivers, ground-handling staff, pilots and the like, it is reasonably straightforward to apply. However, as we move towards autonomous transport and electronic provision of services, this may become more difficult to interpret. Will the computer running the program that controls the vehicle or aircraft be considered an ‘agent, and servant…or any other persons’? It seems unlikely.

Often, an error will have a human origin of some nature – in either the programming or the maintenance of the computer or system. This raises the question: to what extent can a programmer or an IT maintenance employee be considered such an agent, servant or other person for the purposes of the Conventions?

A major concern in the shipping industry is the risk of cyber criminals/hackers gaining control of vessels. Even today with a crew on board, there have been instances of hackers being able to disable vessels and hold them to ransom. The risk multiplies as the physical control on board a vessel reduces (i.e. by having fewer crew on board) - to a point where it is possible that remote operators are locked out, or hackers simply take control of an entirely automated vessel. No doubt there would be many ways that a criminal would take advantage of this. However, the most obvious examples are theft of the vessel (or more likely the cargo), holding it to ransom and, most concerning, use as a weapon.


In an environment where vehicles are increasingly driven by data and technology, rather than by people, there needs to be a clear call-to-action on governments to create modern legal frameworks. These must provide appropriate protocols on the behaviours of vehicle technology, as well as the storage, usage and sharing of the masses of data which will be collected by the next generation of autonomous vehicles. This will require a collaborative approach across government, vehicle manufacturers, software developers, insurers, law enforcement agencies and consumer groups.

This article was first published by Insurance Day on 30 September 2019.

Related articles