- within Transport topic(s)
- in United Kingdom
- with readers working within the Business & Consumer Services and Securities & Investment industries
- within Transport, Food, Drugs, Healthcare, Life Sciences and Environment topic(s)
Autonomous driving: What are self-driving vehicles and what is – legally speaking – the definition of automated driving?
STARTING POINT: What is automated or autonomous driving?
Self-driving (automated) vehicles have been in use in Switzerland for longer than one might assume. As early as 2016, Swiss Post launched a pilot project with self-driving buses. From 2019 to 2021, anyone wanting to travel from the Matte to the Marzili district in Bern could use self-driving vehicles as part of a pilot project run by BERNMOBIL. Currently, SBB, the cantons of Zurich and Aargau, and the NGO Swiss Transit Lab (STL) are jointly implementing another pilot project in Zurich's Furttal district where passengers in Furttal can be picked up by automated vehicles and taken to the nearby S-Bahn station. While still in its test phase, the project aims to begin operational service in the first half of 2026.
Technical support in road traffic is steadily increasing, making autonomous vehicles a part of our new reality rather than a distant dream. In "automated driving," as it is called in legal terminology, the vehicle independently performs certain driving tasks. According to the classification of the Society of Automotive Engineers (SAE), which is also used by the Federal Roads Office FEDRO, automation is divided into five levels (with an upstream level 0).
- Level 0 – not automated
- Level 1 – Assisted
- Level 2 – Partially automated
- Level 3 – Conditionally automated
- Level 4 – Highly automated
- Level 5 – Fully automated
Level 0 – Non-automated:
The vehicle has only warning or assistance functions, such as
automatic emergency braking, blind spot warning, or lane departure
warning.
Level 1 – Assisted driving:
Assistance systems provide either steering OR brake/acceleration
support (lane entering OR adaptive cruise control). The driver
retains full control at all times and must constantly monitor the
system.
Level 2 – Partially automated
driving:
The vehicle combines steering AND brake/acceleration support (lane
entering AND adaptive cruise control at the same time). The driver
remains responsible and must be prepared to intervene at any
time.
Level 3 – Conditionally automated
driving:
The vehicle handles all driving tasks within defined scenarios,
e.g., on highways (e.g., traffic jam assist). Drivers can sit back
but must be ready to take back control when requested.
Level 4 – Highly automated driving:
The vehicle drives completely autonomously on certain routes, such
as in parking garages or specially designated zones (e.g., local
self-driving taxis). The system is capable of independently
establishing safe traffic conditions. Even in an emergency, human
intervention is no longer necessary.
Level 5 – Fully automated driving:
The vehicle independently handles all traffic situations everywhere
(i.e., not just in specific zones as in Level 4) independently.
Human intervention is not required.
The terms autonomous driving or self-driving vehicles typically refer to SAE Levels 4 and 5 (highly automated or fully automated driving).
LEGAL SITUATION: Which legal issues are particularly interesting for automated driving?
Although automated driving has been a topic of discussion in Switzerland for some time, regulation had been sporadic but this changed at the beginning of 2025 when the Road Traffic Act was amended and the federal government enacted the specially designed Ordinance on Automated Driving (OAD). So, it is time to take a look at the (new) regulatory requirements for automated driving in Switzerland.
As vehicles and their electronics become more autonomous and drivers assume fewer and fewer tasks, questions of liability become more pressing. What are the relevant grounds for liability and how will roles be redistributed in automated driving?
Automated driving relies on the appropriate use and processing of data which also raises questions of data protection and cybersecurity. What data may and must be collected, and where can it be stored? How can vehicles and their operating systems be protected against cyberattacks?
In separate blog posts, we will address the topics of regulation, liability, and data protection in with the context of automated driving.
FACTS: An example
To better illustrate the specific issues for interested readers, the blog posts are accompanied by a fictional scenario. The following – fictitious – story is intended to highlight the multitude of (legal) issues arising from automated driving.
The taxi and the red light
Innovate Mobility AG, an up-and-coming Swiss start-up based in Zug, plans to revolutionize passenger transport in urban centers. Its business model is based on a fleet of automated shuttles that can be booked via an app for trips within defined urban zones.
Innovate Mobility AG is cooperating with two specialized partners to implement this project. The vehicles, highly automated SAE Level 4 electric shuttles (model "Sentinel Pod"), are produced by Sentinel Motors Corp., a leading vehicle manufacturer. The complex control software for automated operation, on the other hand, is developed and maintained by CodeDrive Solutions Inc., a software company. The vehicle fleet is controlled and monitored by an operator in a control center at Innovate Mobility AG. The Sentinel Pod has type approval from a European country neighboring Switzerland.
Linda Punctual, a businesswoman from Zurich, regularly uses Innovate Mobility AG's service for her trips to the train station. On a rainy afternoon, she books a Sentinel Pod. Before the trip, the vehicle performed an "over-the-air" software update centrally installed by CodeDrive Solutions Inc. Shortly after Linda Punctual gets in, the vehicle approaches a large intersection. Due to an error in the update, the software misinterprets the red traffic light signal and the Sentinel Pod drives into the intersection without braking.
A vehicle approaching from the right, driven by Mr. Greening, who has correctly entered on a green light, is unable to swerve out of the way and collides with the side of the Sentinel Pod. Both vehicles suffer considerable damage. Mr. Greening suffers minor whiplash. The vehicle and environmental data recorded in the Sentinel Pod is stored on the servers of Sentinel Motors Corp. in the USA immediately after the incident.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.