Legal AI Systems in the EU’s proposed Artificial Intelligence Act

Sebastian Felix Schwemer*, Letizia Tomada, Tommaso Pasini

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

323 Downloads (Pure)

Abstract

In this paper we examine how human-machine interaction in the
legal sector is suggested to be regulated in the EU’s recently
proposed Artificial Intelligence Act. First, we provide a brief
background and overview of the proposal. Then we turn towards the assessment of high-risk AI systems for the legal tasks as well as the obligations for such AI systems in terms of human-machine interaction. We argue that whereas the proposed definition of AI system is broad, the concrete high-risk area of ‘administration of justice and democratic processes’, despite coming with considerable legal uncertainty, is narrow and unlikely to extent into many uses of legal AI and IA systems. Nonetheless, these regulatory developments may be of great relevance for current and future legal AI and IA systems.
Original languageEnglish
Title of host publicationProceedings of the Second International Workshop on AI and Intelligent Assistance for Legal Professionals in the Digital Workplace (LegalAIIA 2021)
PublisherCEUR
Publication date2021
Pages51-58
Publication statusPublished - 2021
EventSecond International Workshop on AI and Intelligent Assistance for Legal Professionals in the Digital Workplace (LegalAIIA 2021) - Sao Paulo, Brazil
Duration: 21 Jun 202121 Jun 2021

Conference

ConferenceSecond International Workshop on AI and Intelligent Assistance for Legal Professionals in the Digital Workplace (LegalAIIA 2021)
Country/TerritoryBrazil
CitySao Paulo
Period21/06/202121/06/2021
SeriesCEUR Workshop Proceedings
Volume2880
ISSN1613-0073

Cite this