The Ethics Committee of the Federal Ministry of Transport and Infrastructure for Automated and Connected Driving has published its report on 20 June 2017, just as the new Road Traffic Act on automated and connected driving is coming into effect on 21 June 2017. Read here > (in German)
The report provides "20 Guidelines" on how to programme automated driving systems. The Committee is an interdisciplinary body of highly reputed legal scholars, scientists, industry and church representatives and other stakeholders. It was established in September 2016, and is led by the former judge at the Federal Constitutional Court, Prof. Dr. Dr. Udo Di Fabio. Read here > (in German)
Non-binding, but relevant – focus on the protection of the human being
The Committee’s Guidelines are non-binding. They are recommendations for the industry (in particular, operators and car manufacturers of the driving systems) to give shape to the requirements of programming automated driving systems. At the same time, they could potentially provide orientation for German courts with regard to liability issues. The Ethics Committee recommends creating a monitoring body that analyses road traffic incidents based on autonomous system failures, as well as a Federal Agency for the safety of automated and connected drive.
The Committee points out that automated driving systems shall, above all, serve the safety of all road users. The protection of humans shall take precedence over any other considerations concerning the usefulness of such driving systems. In case of inevitable accident situations, the driving system shall in no circumstances prioritize the persons involved based on their personal characteristics (such as age, sex, physical or mental constitution).
Guidelines in summary
The following outlines the key aspects of the Guidelines:
- Safety first: Automated driving systems shall, in particular, serve to improve the safety of the road users. They shall increase mobility opportunities and allow further benefits (e.g. greater comfort, physical and mental relief, gain in time).
- Minimize damages: The protection of humans precedes any other reasoning behind such driving systems. The objective is to minimize or avoid damages.
- State control: The admission and control of automated driving systems lie with the public authorities. The admission is acceptable only if the driving systems promise a reduction of damages compared to human driving performance. The avoidance of accidents is a guiding principle, however inevitable remainder risks shall not preclude the introduction of such driving systems.
- Avoiding accidents: The technology shall avoid accidents as much as practically possible, including in so-called dilemma situations. In a dilemma situation the automated vehicle must decide which out of two equivalent hazards will materialize. To that end, manufacturers shall use and further develop all technical possibilities, such as sensor technology, brake performance and “smart” road infrastructure.
- No qualification or quantification of people involved: In case of inevitable accident situations, any qualification or prioritization of people according to their characteristics is prohibited. Any qualification and set-off between the number of potentially affected victims is prohibited. However, manufacturers shall programme the automated system in a fashion to generally reduce the number of persons impaired.
- Liability: The responsibility in the context of automated driving systems shifts from drivers to manufacturers and operators of the technical systems and the infrastructural, political and legal decision-makers. Liability for damages caused by activated automated driving systems shall be subject to the same rules as product liability.
- Information of the public: Operators and manufacturers shall inform the public of new technologies, including their implementation and the ways they are programmed, in a transparent way.
- Data protection: Business models using data that arise out of automated and connected driving systems and vehicle control systems must respect the road users’ autonomy and data sovereignty. Generally, the decision of sharing and using vehicle data lies with the car owners and/or users. A free decision of disclosing the data requires a real alternative and workability for the users.
- "Overruling": The driving system must clearly indicate whether the driver uses the driverless system or whether he remains in firm control of the car ("overruling"). In the latter case, It must be clear and transparent at any time whether the driver or the machine is responsible for / in control of the vehicle (allocation of responsibility). This holds true, in particular, for the interface and transfer of control over the vehicle from the driver to the machine (and vice versa). Such processes, including protocolling them, must follow international standards as yet to be developed.
- Hand-over of control: Highly automated systems shall be programmed in a manner that excludes an abrupt emergency hand-over of control to the driver. In that context, technology must adopt to human behaviour, not the other way around.
- Autonomous vehicles must be able to stop without human intervention in emergency situations.
The German Federal Government is determined to pave the way for automated and connected cars in Germany and to take a lead role in shaping the framework of the future. The work of the Ethics Committee is an integral part of these efforts.