Consultation on the AI Liability Directive Timetable

Today, we start our first public consultation to ask you for insight when it comes to liability scenarios and AI. This is the part of our Better Regulation initiative.

As indicated in my LinkedIn post that informed you about our timetable for the AI Liability Directive (AILD) (link), our first stakeholder consultation starts today. You can send us since 12h00 your contributions via Google Forms (see link below) over the next six weeks until 17 March 2025 18h00.

While Axel Voss and I are still in the middle of our research phase and will continue to read studies and position papers, we have already identified several common statements and observations. The materials have so far indicated that ...

▪️ Liability was from the beginning one of the key fields of the EU AI strategy and the EU Institutions have presented lots of evidence that the special characteristics of AI could create various legal gaps and loopholes.

▪️ The updated PLD does neither seem to cover all AI-related liability scenarios nor all legal actors along the AI value chain. It therefore depends on the Member States laws if the harms / damages are compensated. Even the PLD rules will be implemented very differently across the EU since the law is a directive. The lack of harmonisation in liability creates significant legal uncertainty and presents a huge (relative) competitive disadvantage for SMEs and start ups as those will struggle most under legal / litigation costs.

▪️ It is unfortunate that the AILD was not streamlined more with the AIAct. Not only the definitions and AI operators but also the risk-based approach could have been integrated more in the AILD, something many Member States have repeatedly requested in the Council WG sessions over the past months.

▪️ Since there is no joined and several liability regime in the AILD, downstream companies (= providers or deployers) face the scenario of being left alone with the liability risks. Dominant market players could shift that risk downstream, for instance by forcing their clients to accept a waiver of recourse clause. A potential time bomb in terms of EU competitiveness that is currently regular topic at AI conferences across the EU. 

It is important to note that those four points are all not fully verified yet as we continue to study the status quo. Having the Draghi report in mind and being aware of the strong regulatory burden for EU companies, we will only move forward if we feel that the AILD will reduce legal uncertainty & fragmentation, strengthen our economy and competitiveness, and help claimants (= can be both, a company or an individual) to get compensated. Please help us on our legislative journey with your input and insights 📚 🤖 💡

Access the consultation here.