Your Guide to Explaining Autonomous Vehicles Regulations in Europe

Are autonomous vehicles more dangerous than a human behind the wheel? 

Can robo-taxis replace ride-hailing without raising accident rates on our streets? 

Will driverless cars ever move into the mainstream? 

These are some of the questions operators, regulators and insurers are asking themselves in  2026.

Waymo recently waded into the debate by publishing a report on driverless cars and accident rates. Their research showed fewer crashes during trips than those related to human drivers. A staggering 68% fewer crashes than a person would typically be in. 

The data relates to the operations in the US and their adoption with autonomous vehicles, but it poses the question where is Europe with the progress of automated vehicles?

How are autonomous vehicles defined in Europe?

Before we dive into where each country is with their progress on autonomous vehicles (AV). Let’s talk about levels of autonomous driving. These set the foundation for the degrees of regulation to come. 

SAE has published the framework for levels of driving automation. It describes different levels from 0 meaning no driving automation to 5 having full driving automation. 

Going more in-depth. Levels 0 to 2 mean that there’s human driver responsibility. Level 0 meaning that humans are fully in control and doing all the driving, steering, and braking. 

Level 1 means having driver assistance, the system helps with either steering or speed control. The next level from that means the system controls both steering and speed at the same time. 

Level 3-5 means the system is more responsible and vehicles are being more automated. 

Level 3 meaning that the car can drive itself under specific conditions and monitor the road. In that case humans must be ready to take over the car when needed. 

The difference between level 4 and 5 is that one is just highly automated and the other one is fully autonomous.
Therefore, level 5 means that the car is able to drive itself everywhere, in any weather or location without needing a human. 

How does Europe regulate autonomous vehicles?

In Europe, there’s no licensing practice in the EU that allows fully autonomous vehicles (level 5). Countries have been adopting regulations at their own pace with various degrees of specifications relating to the level of autonomous capability. 

Recently, the United Nations Economic Commission for Europe (UNECE) adopted the first global rules for allowing fully autonomous driving systems (ADS) (SAE levels 3 and 4). The framework was made on June 24, 2026 and is expected to enter into force one month later.

These requirements are designed to build public trust and confidence that AVs are safe to deploy. Below we break down the current state of play country by country where regulation or pilots currently exists. 

Germany 🇩🇪

In Germany it’s already legal to drive with autonomous vehicles. 

They published the world's first comprehensive level 4 law in 2021 and 2022, allowing driverless vehicles on public roads within defined operational design. Setting strict safety standards that require AVs to have black-box records. The liability goes with a ‘technical supervisor’ who monitors multiple AVs and can interrupt.

UK 🇬🇧

Has one of the most detailed national frameworks worldwide, covering approval, authorisation, operator licensing and in-use regulation. They have a ‘user-in-charge’ concept that means companies need to show that AVs are safer or even safer than human drivers. The liability is placed on the self-driving entity rather than the user.

France 🇫🇷

France’s regulatory status is developing operational framework and pilot enabling. 

France has made adjustments to enable AVs including level 4 automated vehicles to be tested and operated under controlled conditions. 

Regulatory focus is on experimental deployments, data sharing and safety demonstrations.

Netherlands 🇳🇱

Netherlands has an extensive testing framework. They have implemented a dedicated legal framework for AV testing, permitting on road trials with vehicle authority and including driverless tests in controlled scenarios. 

It’s a leading jurisdiction in the EU for structured AV experiments. Next is to expand permitted conditions and integrate with broader EU testing harmonisation efforts. 

Finland 🇫🇮

They’re in the stage level 4 testing, while most of the countries besides Germany are in level 3. 

Finland has their framework available and enabling laws is in progress. They allow experimentation with automated vehicles, supported by recent national laws that further enable AV activities. 

They follow EU and UNECE type-approval requirements, and are working on specific implementation rules and broader on-road test authorisations that are part of their legislative development. 

How does insurance fit in it?

Before operating on the roads all autonomous vehicles need to carry valid motor third-party liability (MTPL) insurance. Coverage extends across EU member states, and MTPL exists to protect victims. Injured third parties or damaged property are compensated directly through MTPL claims. 

But autonomous vehicles are quietly breaking this logic. With these vehicles the liability is changing from a human to technology provider behind the system. 

Traditional insurance is not designed for this shift.

When an accident happens with an automated driving system engaged, the MTPL insurer still pays the victim first. Just like it was before. Only afterwards they have a legal right to pursue the vehicle manufacturer or software developer, if a clear product or software failure caused the crash. 

That’s a different claims process that insurers are used to.

It gets even more complicated in mixed scenarios, where control switches between driver and system during a single trip. Autonomous vehicles need to log their driving mode data to assign the fault correctly. Whether it was the human or the system who was in control. 

Adding another layer to that, when the owner has skipped critical safety updates or ran unauthorised modifications, policies often limit insurer liability. 

All of this means that the claims process needs new capabilities of reading operational data, interpreting driving mode logs, and assigning liability with accuracy that traditional risk carriers were never built for. 

Insurers are starting to explore system-centric policies that treat the vehicle less like a car and more like a mobile software platform.  This is where modern insurance solutions matter. AV companies need partners who can support on the technical side well enough to structure policies around the system behaviour, not just driver history. 

This article is for general information only and does not constitute legal advice.