The rampant success of Generative AI services poses a new copyright conundrum. All over the world, it is discussed whether the training of General-Purpose AI models (GPAIM) with copyright protected works for commercial purposes falls within the exclusive bundle of rights of copyright holders. In the EU, some last-minute provisions have been introduced in the recent AI Act, pointing towards the application of the text and data mining exceptions set forth by the directive 790/2019/EU (CDSM directive) to solve the issue despite the fact that Generative AI was not considered by the European legislator when implementing the exceptions in 2019. According to Art. 4.3 of the CDSM directive, copyright holders can opt-out of the exception and thus reserve their content for the extraction of their intellectual works to improve machine learning performances. To monitor the respect of the opt-out, the AI Act introduces the obligation to draw up and make publicly available a sufficiently detailed summary about the content used for training of the GPAIM. Providers of generative AI services should abide by the mentioned transparency obligation by filling a template provided by the AI Office, which is the new authority established by the European Commission to ensure the implementation of the AI Act. As regards to copyright matters, the AI Office, apart from issuing the mentioned template, has the power to make recommendations and to monitor the enforcement of the regulation. This paper compares the foreseeable promises and limitations of the AI Office to ensure sustainable creativity in Europe after the tectonic shift brought by Generative AI to the already fragile balance of the copyright system. It argues that the complexity of copyright issues in the digital environment should invite to a wider reflection on the proper institutional framework regulating EU copyright law, calling for a pan European structure which fosters coordination among the new administrative bodies set up in the recently sectorial interventions dealing with digital issues, especially the Digital Services Act, the Digital Markets Act, the Data Governance Act and the European Media Freedom Act.

Towards an Independent EU Regulator for Copyright Issues of Generative AI: What Role for the AI Office (But More Importantly: What's Next)? / Geiger, Christophe; Iaia, Vincenzo. - In: AUTEURS & MÉDIA. - ISSN 1370-6039. - 2(2024), pp. 188-199.

Towards an Independent EU Regulator for Copyright Issues of Generative AI: What Role for the AI Office (But More Importantly: What's Next)?

Geiger, C.
;
Iaia, V.
2024

Abstract

The rampant success of Generative AI services poses a new copyright conundrum. All over the world, it is discussed whether the training of General-Purpose AI models (GPAIM) with copyright protected works for commercial purposes falls within the exclusive bundle of rights of copyright holders. In the EU, some last-minute provisions have been introduced in the recent AI Act, pointing towards the application of the text and data mining exceptions set forth by the directive 790/2019/EU (CDSM directive) to solve the issue despite the fact that Generative AI was not considered by the European legislator when implementing the exceptions in 2019. According to Art. 4.3 of the CDSM directive, copyright holders can opt-out of the exception and thus reserve their content for the extraction of their intellectual works to improve machine learning performances. To monitor the respect of the opt-out, the AI Act introduces the obligation to draw up and make publicly available a sufficiently detailed summary about the content used for training of the GPAIM. Providers of generative AI services should abide by the mentioned transparency obligation by filling a template provided by the AI Office, which is the new authority established by the European Commission to ensure the implementation of the AI Act. As regards to copyright matters, the AI Office, apart from issuing the mentioned template, has the power to make recommendations and to monitor the enforcement of the regulation. This paper compares the foreseeable promises and limitations of the AI Office to ensure sustainable creativity in Europe after the tectonic shift brought by Generative AI to the already fragile balance of the copyright system. It argues that the complexity of copyright issues in the digital environment should invite to a wider reflection on the proper institutional framework regulating EU copyright law, calling for a pan European structure which fosters coordination among the new administrative bodies set up in the recently sectorial interventions dealing with digital issues, especially the Digital Services Act, the Digital Markets Act, the Data Governance Act and the European Media Freedom Act.
2024
Towards an Independent EU Regulator for Copyright Issues of Generative AI: What Role for the AI Office (But More Importantly: What's Next)? / Geiger, Christophe; Iaia, Vincenzo. - In: AUTEURS & MÉDIA. - ISSN 1370-6039. - 2(2024), pp. 188-199.
File in questo prodotto:
File Dimensione Formato  
Geiger_Auteurs et Média_compressed_Parte1.pdf

Solo gestori archivio

Tipologia: Versione dell'editore
Licenza: Tutti i diritti riservati
Dimensione 9.95 MB
Formato Adobe PDF
9.95 MB Adobe PDF   Visualizza/Apri
Geiger_Auteurs et Média_compressed_Parte2.pdf

Solo gestori archivio

Tipologia: Versione dell'editore
Licenza: Tutti i diritti riservati
Dimensione 8.09 MB
Formato Adobe PDF
8.09 MB Adobe PDF   Visualizza/Apri
Geiger_Auteurs et Média_compressed_Parte3.pdf

Solo gestori archivio

Tipologia: Versione dell'editore
Licenza: Tutti i diritti riservati
Dimensione 7.92 MB
Formato Adobe PDF
7.92 MB Adobe PDF   Visualizza/Apri
Geiger_Auteurs et Média_compressed_Parte4.pdf

Solo gestori archivio

Tipologia: Versione dell'editore
Licenza: Tutti i diritti riservati
Dimensione 3.71 MB
Formato Adobe PDF
3.71 MB Adobe PDF   Visualizza/Apri
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11385/244738
Citazioni
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
  • OpenAlex ND
social impact