By Ben Howarth, senior adviser for motor and liability, Association of British Insurers (ABI)

Most estimates say that nine in every 10 crashes on UK roads involve some element of human error, so it would seem to follow that automating the driving task will reduce the overall number of crashes.

As an industry, we have been really encouraged by the proactive approach taken by the Government’s Centre for Connected and Autonomous Vehicles (CCAV) and we are confident this will help deliver what consumers need.

Given the economies of scale, it is likely commercial fleets will be among the first pioneers to use truly automated vehicles on our roads.

Legislation was introduced to Parliament earlier this year to clarify how insurance will work for the first wave of fully automated vehicles.

The insurance industry strongly supported these proposals, which struck the right balance between incentivising the uptake of this technology without disrupting the well-established and understood process of bringing an insurance claim.

The legislation, unfortunately, did not make it into law before the Prime Minister called June’s snap general election.

But, the Government quickly re-committed to bringing the proposals back in a new Bill (now called the Automated Driving and Electric Vehicles Bill), which we expect to be formally introduced in the autumn.

As the proposals had previously enjoyed cross-party support, we are confident they will become law and that means insurers are now getting ready for the first of these vehicles to come onto the road.

We still don’t know exactly when that will happen.

The technology itself is nearly ready.

It has, after all, been 12 years since a team from Stanford University won a $2 million (£1.5m) prize for successfully sending ‘Stanley’, a fully self-driving car, on a complex course through a Californian desert.

But there is much to do to persuade the public.

A recent survey from Lexis Nexis showed only half of the public think driverless cars are a good idea.

A minority (14%) don't think there will be any benefits whatsoever.

A crucial commitment made by the Government is that it will take responsibility for defining what is and isn’t automated driving.

Although the details need to be worked through, this is a very welcome commitment and offers reassurance that people will not be misled into thinking a car is ‘driverless’ when, in reality, it relies on the driver to supervise the system and intervene if needed.

That is the reason why the ABI and Thatcham Research have recently published a detailed proposal for the safety frameworks that should underpin automated driving.

Although individual manufacturers will want to differentiate themselves from their competitors, drivers will need to understand the distinction between ‘assisted’ driving (where technology makes the driving task easier, but never takes overall control of the vehicle) and ‘automated’ driving (where the technology performs every part of the driving task).

In particular, it is vital that there are redundancy systems in place so that, while in automated mode, emergency manoeuvres can be performed without the driver needing to make a split-second intervention.

We are confident that these safeguards will make the driving public feel much more confident about using this technology.

Like the Government, the insurance industry is treating the development of automated driving as a strategic priority.

We recognise the model for settling insurance claims will need to be adapted in line with technological innovation.

Insurance law needs to ensure that, when a car is in automated mode, the ‘driver’ cannot be blamed for an accident they could not reasonably be expected to prevent.

In those circumstances, the ‘driver’ would often be the victim, entitled to compensation themselves.

That does not mean the Government needs to re-write the whole Road Traffic Act.

After all, the automated mode will initially only be included in a small number of top-end vehicles and, even then, only used in the safest driving situations.

Instead, we aim to adapt the existing motor insurance system, while working with vehicle manufacturers to establish a system for the costs of these claims to be apportioned fairly.

Under our simple system, the insurer will always settle the claim, as they do today, but will then have a right of recovery against a manufacturer for accidents occurring while the vehicle was in automated mode.

Although the long-term vision might be a seamless network of connected technologies, the reality is likely to be somewhat messier as different technologies serving different needs compete for the public's attention.

Insurance will be crucial to managing risks as it is clear not every driver will be moving along the ‘pathway to driverless cars' at the same speed.

There are questions for the industry to grapple with, such as will insurers rely on the courts to recover their costs from manufacturers or will the two industries collaborate directly?

Will responsibility for ensuring the correct insurance cover is in place fall on individuals or on the manufacturers selling the cars?

How will insurers respond if it becomes clear that certain models of automated driving have a worse safety performance than others?

Will an after-market develop for these vehicles and who will be responsible for ensuring technology and software is up-to-date?

For these reasons, many insurers expect commercial fleets to be the first part of the overall vehicle parc to adopt automated driving technology in significant numbers.

While many of these issues will remain complex, there will be economic incentives for fleet operators, insurers and manufacturers to work together to address them.

I would expect to see insurers taking the lessons they learn from early adopters in the fleet sector and then applying these as automated driving becomes more common in the private car market.

Fleet operators will also be well aware that, if they are to be able to make best use of automated technology, they need the wider driving public to be confident about being on the roads at the same time.

A priority is ensuring that there is absolutely no ambiguity about when a car was operating autonomously and when it wasn’t, especially after an accident.

There must be an adequate data-sharing system to ensure all parties can immediately establish whether an automated mode was enabled at the time of a collision and if it was being used correctly.

The ABI has worked with its members to set out clearly the data a claims handler will need to have readily available when a claim is brought to help establish this.

We have had encouraging signals from manufacturers recently, but we will be pressing all stakeholders to make sure this process is finalised before vehicles are licensed for use on the road.

Over the past decade, the insurance industry has worked increasingly to build a positive inter-dependence with vehicle and fleet operators, as the industry’s focus has shifted from being purely about financial risk transfer to being about preventing claims in the first place.

Automated driving technology is the logical conclusion to work over several decades to reduce the numbers of people killed or seriously injured on the roads.

The first part of the transition to automated driving – setting a clear liability framework for the first adopters of this technology – is almost complete.

A bigger challenge, which all parts of the road-using economy need to work together on, is agreeing exactly when and how that technology can be used.

The ABI will be focussing on this question over the coming months and years, and we look forward to working closely with our partners in the fleet industry to make automated driving a success.