In a study published on Thursday 19 September, the European Parliamentary Research Service (EPRS) recommends broadening the scope of the future ‘AILD’ directive on adapting civil liability rules to artificial intelligence.
This directive, proposed by the Commission in 2022, aims to incorporate a liability system for civil damages created by AI systems. It goes hand in hand with the changes made to the directive on liability for defective products (see EUROPE 13314/15), but completes its gaps on specific AI risks, such as discrimination, hate speech and fundamental rights.
The study recommends that the directive should cover general-purpose AI as well as its prohibited and high-risk uses, as already defined in the AI Act.
The EPRS also advocates transforming the proposed AILD into a broader instrument on software liability “to prevent market fragmentation and enhance clarity across the EU”.
Critically, the study argues that the Commission should take account of the Parliament’s 2020 resolution, which calls for a “strict liability” regime to be imposed on operators of high-risk AI systems, implying legal liability even in the absence of fault or criminal intent.
The AILD is still awaiting scrutiny by parliamentary committees, particularly the Committee on Legal Affairs, which is due to resume its work by the end of October.
See the study: https://aeur.eu/f/djk (Original version in French by Isalia Stieffatre)