Garbage In, Garbage Out. Regulating Generative AI Through Copyright Law

ZUM 10/2023, 645-660

22 Pages Posted: 13 Oct 2023

See all articles by Katharina de la Durantaye

Katharina de la Durantaye

Humboldt University of Berlin - Faculty of Law

Date Written: August 28, 2023

Abstract

Both in the U.S. and in the EU, it is a matter of controversy whether copyrighted content may be used to train generative AI. In the U.S., the discourse circles around the impact which copyright has on innovation and competition. In the EU, practical problems dominate the discussions. These problems have led rights holders to formulate concrete policy demands. The European Parliament wants to address these demands in the draft AI Act (AIA). The rule it proposes (Art. 28(4)(c)) threatens to further diminish the extent to which EU Member States participate in the value creation which the development of generative AI entails. As things are, European copyright law already risks to undermine the AIA’s provisions on product safety. Instead of focusing on practical problems, we should devote our attention to how we want to structure the societal change which generative AI will bring about, and on the role which we want copyright to play in this change.

Keywords: Copyright, generative AI, AI Act, AIA, TDM, text and data mining, fair use, European Union, EU, United States, U.S.

JEL Classification: K21, K29, K39

Suggested Citation

de la Durantaye, Katharina, Garbage In, Garbage Out. Regulating Generative AI Through Copyright Law (August 28, 2023). ZUM 10/2023, 645-660, Available at SSRN: https://ssrn.com/abstract=4572952 or http://dx.doi.org/10.2139/ssrn.4572952

Katharina De la Durantaye (Contact Author)

Humboldt University of Berlin - Faculty of Law ( email )

Unter den Linden 6
Berlin, D-10099
Germany

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
554
Abstract Views
1,241
Rank
100,751
PlumX Metrics