Autonomous Vehicles & Insurance: Who’s Liable When No One’s Driving?

The rise of autonomous vehicles isn’t just transforming roads, it’s reshaping responsibility. As cars begin to drive themselves, a major question arises: who’s at fault when something goes wrong?

For centuries, motor insurance has centred around driver behaviour. But in the age of automation, we’re seeing a seismic shift. Governments, insurers, and manufacturers are now grappling with an unprecedented reality: what happens when no one is holding the wheel?

In this RiskWire feature, we dive into how insurance is evolving for autonomous vehicles in the UK and EU, who will carry the blame and the bill in the event of a crash, and what the future holds for underwriting, claims and consumer trust.

UK: Leading the Liability Shift. 🇬🇧

The UK has emerged as a global front-runner in defining self-driving liability. The Automated and Electric Vehicles Act 2018 laid the groundwork by introducing the “single insurer model” where the insurer pays out, then recovers from the tech provider if needed.

Fast-forward to 2024, and the Automated Vehicles Act has gone further:

This means that during autonomous operation, the human is just a passenger. The authorised self-driving entity typically the manufacturer or software provider carries the risk. It’s a bold move, but one insurers broadly support, clarity creates confidence.

EU: Cautious But Catching Up. 🇪🇺

Across the Channel, the EU is still piecing together its approach. Current frameworks rely on the Motor Insurance Directive and the Product Liability Directive, both of which pre-date AVs.

Progress is happening. Draft proposals for an AI Liability Directive and updates to product liability laws aim to streamline compensation and clarify who’s responsible when intelligent systems cause harm.

Some Member States have gone further:

• Germany 🇩🇪 has legal definitions for Levels 3–4 automation and places some liability on manufacturers.
• France 🇫🇷 has modified its laws to cover autonomous operation under its no-fault traffic compensation rules.

But unlike the UK’s unified AV law, the EU remains a mosaic, for now.

Who’s Really at Fault? It Depends.

Let’s break it down:

• Level 2–3 Automation: The human still has a duty to monitor and intervene. If they don’t, they can still be found liable.
• Level 4+: The vehicle is in charge. In these cases, liability sits with the tech provider or manufacturer, especially in the UK.

Notably, Mercedes-Benz and Volvo have both stated they will accept full liability if their certified systems are at fault while driving autonomously.

This shift from personal to product liability is huge. And it’s forcing insurers to rethink the entire model.

How Insurers Are Responding. 🛟

Insurers in the UK and EU are actively preparing for this transition:

• New product structures: Policies are evolving to cover both manual and autonomous driving under a single premium, with the insurer recovering costs from the manufacturer if needed.
• Trial participation: Firms like Direct Line, AXA, and Aviva are directly involved in AV pilots insuring autonomous shuttles, robotaxis, and more.
• Specialist cover: AXA XL has developed policies that include cyber risk, software failure, and trial interruption for AV developers.
• Data-driven pricing: The rise of sensors, ADAS features and black box data is shifting the focus from driver history to vehicle behaviour.

What’s clear? The role of insurance is moving from personal protection to technological accountability.

Examples in Action.

• Direct Line x FiveAI: Insured one of the UK’s first suburban AV trials.
• Apollo x Oxbotica: Offered bespoke cover for Europe’s first zero-occupancy public AV journey.
• AXA x Navya: Partnered to insure driverless shuttles across France and beyond.
• Aviva: Covered a Cambridge AV shuttle with no onboard driver one of the UK’s earliest.

These are more than pilots they’re blueprints for future insurance models.

The Road Ahead: What to Expect by 2030. 🛣️

• Liability will shift upstream: Manufacturers will carry more risk; personal motor insurance may shrink.
• Insurance will be embedded: Automakers could bundle coverage into vehicle sales or mobility services.
• Fleet-first future: Commercial AV insurance (for robotaxis, delivery pods, etc.) will become a major market.
• Cyber + AI cover will rise: As tech risks grow, so will demand for specialist insurance data loss, hacking, software bugs, and more.
• Premiums may fall: Fewer crashes, better risk profiles AVs could mean cheaper insurance overall.

But there’s a catch: we’re entering a “messy middle”. For the next 5–10 years, roads will be a mix of human and machine drivers. That makes underwriting riskier, not easier at least for now.

Final Thoughts.

Autonomous vehicles are not just a technology challenge they’re a legal, ethical and insurance challenge.

The UK is showing what’s possible: clear laws, defined liability, and insurer readiness. The EU is getting there, but more harmonisation is needed.

For insurers, this isn’t a threat it’s an opportunity. A chance to lead the transition, design future proof products, and build public trust in a world where humans are no longer in the driving seat.

Written by The RiskWire Team. ✌️