Liability for Artificial Intelligence and Software

The study on ‘Civil Liability for Artificial Intelligence and Software’, initiated by the European Commission, analyses how damage caused by artificial intelligence (AI) systems and by software is allocated by the rules of tortious liability currently in force in the EU, and whether – and, if so, to what extent – national tort law regimes differ in that respect. It also examines possible gaps in the protection of the injured parties. The study assesses solutions expected on the basis of current European tort laws after weighing up conflicting interests, and goes on to identify questions in respect of which the interpretation of the liability rules in force appears uncertain as well as potential inconsistencies in said solutions. It focuses on the key aspects of selected, most significant problem constellations and uses illustrative examples from selected European legal systems representing different legal families. The project also covers the American perspective.

The study is conducted by ETL in cooperation with ECTIL and is led by Ernst Karner (ETL, ECTIL and Vienna University) and Bernhard A Koch (Innsbruck University). The liability for software is presented by Christiane Wendehorst (Vienna University), and the part covering American law is written by Mark Geistfeld (New York University School of Law).

The study began in the spring of 2020 and was completed as regards content in the first half of 2021.

A publication of the project results in English is planned for mid-2022 with De Gruyter (Berlin/Boston).