By Michael Hookham, Head of Commercial Motor.
Recently at the supermarket there was a dog sitting in the driver's seat of a car. When the driver got in obviously the dog moved over – but in a world of automtous vehicles will this change? This is a question the insurance industry is quietly grappling with right now: as cars take on more of the driving themselves, who is in control? And when something goes wrong, who's responsible?
Built on human error
Motor insurance is built around one simple concept: humans make mistakes, and its these mistakes that while are not ideal, are accepted. We forgive it, price for it, and the whole industry sits on that foundation. But the technology that was optional - or didn't exist - in cars twenty years ago is now standard: cameras, sensors, automatic braking, lane assistance. As I put it to the room: the moment a car itself starts making the decision, it opens up a genuinely different question of who's responsible for that decision - the car, the owner, the retailer, the software developer, the hardware manufacturer, or whoever's kept it maintained. Nobody in the industry has a complete answer yet as it has not yet been tested.
It's not just the car
The vehicle is only one part of the picture. Autonomous systems depend on reading the world around them correctly - road markings, signage, infrastructure. If a street sign is graffitied and a self-driving car can't read it properly, and that leads to a crash, who's responsible: whoever damaged the sign, or the council responsible for maintaining it, the car or the driver for not intervening? These are the kinds of cases that will increasingly show up in claims.
There's a harder layer too. In a genuinely unavoidable accident, an autonomous vehicle may need to make a split-second judgement call about outcomes - a decision that today is still left to human instinct in the moment.
The trouble is, instinct doesn't always cooperate, even with existing technology. Anti-lock braking is a good example: it's designed to work if you keep firm, steady pressure on the brake pedal while the system pulses it faster than any human could. But many drivers panic and take their foot off the brake entirely - the one thing that undoes what the system was built to do.
As vehicles take over more of this kind of decision-making, someone must set the rules for how those judgement calls get made - and that's increasingly a job for governments, not manufacturers. How will it decide which has the least damaging outcome. As countries have difference political, religious, and cultural ideologies these outcomes will shift from country to country based on the Governments different prioritisations. Different countries may land on different answers.
It's also worth noting there's no universal technical standard across manufacturers globally, partly because the underlying systems are closely guarded IP - which only adds to the complexity of getting everyone on the same page.
A global build, not a global standard
Vehicle manufacturing over the last couple of decades has been shifting to countries that can access labour, material, and other aspects to build cars as efficiently as possible through economies of scale. This is shifting away from countries making everything themselves, toward a more globalised model - production centred wherever the resources, labour, and supportive regulation exist. Different regions are also setting different rules for what a vehicle needs to meet, whether emissions, safety, or autonomous driving standards. Vehicles built to different regional rules may not "talk" to each other the same way, adding friction just as the fleet becomes more automated.
A recent Delta claim that tested this exact question
We recently handled a claim that's a small but useful example. A client's vehicle was legally parked, partly on a grass verge. A neighbour's robotic lawnmower - programmed to a fixed area - mowed into an obstacle it wasn't designed to detect: the underside of the parked car. It became lodged there and was damaged when the vehicle was later driven off. The homeowner sought recovery from our client. We successfully defended it: the vehicle was parked reasonably, and it's not reasonable to expect a driver to check underneath their car for a hidden object before moving off. It's a small case, but exactly the kind of dispute - human behaviour meeting an automated system with limited awareness of its surroundings - we'll see far more of as automation increases.
How far away is this, really?
The average age of a vehicle on New Zealand roads is 14.8 years - one of the oldest fleets in the developed world, several years older than Australia's or the UK's - and cost-of-living pressure is likely to push that higher, not lower.¹ Even if every new car sold from today were fully autonomous, I'd estimate at least 20 years for anything approaching a full changeover, and car culture - hobbyists, motorsport, restoration - means traditionally driven vehicles will share our roads with autonomous ones for a long time after that.
There's an upside too: mobility for people who can't currently drive, and cities potentially reclaiming space currently locked up as parking if cars can drop passengers off and park themselves elsewhere.
What brokers can do now
Ask what's already in your clients' vehicles - many newer cars already carry driver-assist or semi-autonomous features and its worth discussing with your clients as to the reduction this may be having on the claim's frequency, however also the impact it may be having on repair costs. Is there information captured from the vehicles that can be used to assist with determining who's liable and any potential recoveries? If a client has put real work into reducing risk - training, telematics, maintenance programmes - make sure that's visible at renewal, not just their claims history. Clients who invest in safety shouldn't end up paying the same as those who haven't.
Where Delta sits
We've built our commercial motor cover with this shift already in mind - including cover for EVs and autonomous driving functionality, and recovery processes that don't ensure other people's mistakes don't impact our clients' premiums. As the technology and legislation evolve, this is a case of refining what we already have, not starting from scratch.
Nobody has all the answers on where liability ultimately lands. But the conversation is already happening in claims files, well before most people expect it to - and we'd rather have it with you now than after it's already changed. If you want to talk through what this means for your clients, get in touch.
¹ Source: MoneyHub NZ, New Zealand Vehicle Fleet Statistics
