The EU expert report on the ethics of self-driving cars – four years later

A nice event on societal dimensions of connected and automated mobility organized by the CCAM and the Polish government in Gdansk. I have talked about the ethical recommendations of our EU expert group from 2021. And what has happened in the meanwhile (with Fabio Fossa).

Gdańsk is a lovely city on the Baltic Sea, famous for being the hometown of the glorious Solidarność trade union of Lech Wałęsa. Beautiful historical centre, nice cafes along the river.

CCAM is a European partnership and an association supporting research, development and policy in the field of connected and automated mobility. There is a State Representative Group (SRG) supporting the initiative, and Poland has the presidency for this year. On April 23rd 2025, CCAM and SRG (Polish government) have organized a CCAM meeting and a preparatory workshop titled “Basic conditions of societal acceptance for AV (Autonomous Vehicles). I was invited to give a talk based on my experience as member and rapporteur of the report of the EU independent expert group on ethical issues with driverless mobility.

I thought it would be a good opportunity not only to present and discuss some ideas from that report but also to reflect on whether the events of the last four years have raised new questions and issues. So I thought I would invite Fabio Fossa from the Politecnico di Milano to join me in the adventure. Fabio is a big expert on the ethics of self-driving car (check his book to believe) and among other things teaches the course in Ethics of Transportation that I had started many years ago as an adjunct professor in Milan.

I presented the main ideas of the 2020 report:

• moving from a “techno-solutionst” narrative (self-driving cars will make road safer, greener, etc.) to a Responsible Innovation one: driving automation may contribute to safety, sustainability if developed in the interest of people and a broad variety of ethical and political values
• The three main areas of ethical attention are: responsible management of risks and safety in the development and testing: data ethics; and responsibility distribution
• Responsibility does not equate legal liability (or moral culpability) for accidents but also, and more importantly, the ability and willingness of different stakeholders to live up to their obligations to ensure a responsible and beneficial introduction of connected and automated mobility

We asked the participants to “vote” about which of these issues they saw as particularly urgent today, and only a minority mentioned data ethics.

Very interesting, since, as Fabio Fossa explained in the second part of our talk, two big things that have happened since the publication of the report have to do with data and AI. These are:

• the AI act, that has categorized self-driving cars as high risk AI application (making ethical and legal considerations even more urgent)

• the explosion of GenAI/LLM (ChatGPT and the like). It is not difficult to imagine its possible uses in, for instance: integrating the functioning of CAVs by filling sensor data gaps, assisting the users with explanations about the CAV’s functioning or supporting manufacturers with virtual training, testing and validation. It is also not difficult to imagine issues with these application in terms of safety, justice, accountability etc.