The European Union Artificial Intelligence Act (EU AIA) contains provisions that would require providers of general purpose AI models to comply with transparency obligation regarding the data using for training their models. It also contains a provision that would require providers of general purpose AI models “to put in place a policy to respect Union copyright law in particular to identify and respect, including through state of the art technologies, the reservations of rights expressed” under the text and data mining (TDM) exception in Article 4(3) of Directive (EU) 2019/79. There is considerable confusion about how to construe this latter requirement, and in particular, whether the requirement to respect EU copyright law is intended to expand the extra-territorial effect of EU copyright laws or to place a trade restriction on general purpose AI models which would require them to comply with EU copyright law in order to be made available in the EU. This blog post examines this question.
The Text of the EU AIA
The relevant provisions of the EU AIA related to respecting EU copyright law are set out below.
The relevant recitals read as follows:
(104) The providers of general purpose AI models that are released under a free and open source license, and whose parameters, including the weights, the information on the model architecture, and the information on model usage, are made publicly available should be subject to exceptions as regards the transparency-related requirements imposed on general purpose AI models, unless they can be considered to present a systemic risk, in which case the circumstance that the model is transparent and accompanied by an open source license should not be considered a sufficient reason to exclude compliance with the obligations under this Regulation.
In any case, given that the release of general purpose AI models under free and open source licence does not necessarily reveal substantial information on the dataset used for the training or fine-tuning of the model and on how thereby the respect of copyright law was ensured, the exception provided for general purpose AI models from compliance with the tranparency-related requirements should not concern the obligation to produce a summary about the content used for model training and the obligation to put in place a policy to respect Union copyright law in particular to identify and respect the reservations of rights expressed pursuant to Article 4(3) of Directive (EU) 2019/79
(105) General-purpose models, in particular large generative models, capable of generating text, images, and other content, present unique innovation opportunities but also challenges to artists, authors, and other creators and the way their creative content is created, distributed, used and consumed. The development and training of such models require access to vast amounts of text, images, videos, and other data. Text and data mining techniques may be used extensively in this context for the retrieval and analysis of such content, which may be protected by copyright and related rights. Any use of copyright protected content requires the authorization of the rightholder concerned unless relevant copyright exceptions and limitations apply. Directive (EU) 2019/790 introduced exceptions and limitations allowing reproductions and extractions of works or other subject matter, for the purposes of text and data mining, under certain conditions. Under these rules, rightholders may choose to reserve their rights over their works or other subject matter to prevent text and data mining, unless this is done for the purposes of scientific research. Where the rights to opt out has been expressly reserved in an appropriate manner, providers of general-purpose AI models need to obtain an authorisation from rightholders if they want to carry out text and data mining over such works.
(106) Providers that place general purpose AI models on the EU market should ensure compliance with the relevant obligations in this Regulation. For this purpose, providers of general purpose AI models should put in place a policy to respect Union law on copyright and related rights, in particular to identify and respect the reservations of rights expressed by rightholders pursuant to Article 4(3) of Directive (EU) 2019/790. Any provider placing a general purpose AI model on the EU market should comply with this obligation, regardless of the jurisdiction in which the copyright-relevant acts underpinning the training of these general purpose AI models take place. This is necessary to ensure a level playing field among providers of general purpose AI models where no provider should be able to gain a competitive advantage in the EU market by applying lower copyright standards than those provided in the Union.
(108) With regard to the obligations imposed on providers of general-purpose AI models to put in place a policy to respect Union copyright law and make publicly available a summary of the content used for the training, the AI Office should monitor whether the provider has fulfilled those obligations without verifying or proceeding to a work-by-work assessment of the training data in terms of copyright compliance. This Regulation does not affect the enforcement of copyright rules as provided for under Union law.
Article 52(1)(c) of the EU AIA states that providers of general purpose AI models shall “put in place a policy to respect Union copyright law in particular to identify and respect, including through state of the art technologies, the reservations of rights expressed pursuant to Article 4(3) of Directive (EU) 2019/790”.
Interpreting the EU AIA requirement to respect EU copyright law
At the recent Fordham IP Conference, I had the opportunity to discuss with leading EU lawyers and academics what the EU AIA policy to respect EU copyright will require.
One view, is that the policy to respect EU copyright law would have the effect of expanding the territoriality of EU copyright law applied to the training if GenAI models. That is, generally the reproduction right is applied on a territorial basis and so training of AI models that takes place outside the EU, e.g. in the United States, would ordinarily not implicate the reproduction right under the copyright laws of EU member states. However, the wording of Recital 60(j) that would require any provider of a GenAI model to comply with EU copyright law “regardless of the jurisdiction in which the copyright-relevant acts underpinning the training of these general purpose AI models take place” could be viewed as extending the scope of EU copyright law.
Experts on copyright I spoke with suggested this interpretation of these provisions could raise questions about compliance with copyright conventions and treaties which generally adopt a territorial approach (the lex loci delicti rule) to copyright and the private international law principle of comity.
A second view, is that the policy to respect EU copyright law in the EU AIA is intended to operate as a trade restriction to prohibit a provider from making a GenAI model available in the EU unless the model respects EU copyright law. This was the view expressed by Dr. Ole Jani, a German copyright lawyer, who answered a question I asked at a terrific panel at Fordham “Artificial Intelligence – Input” moderated by Joshua Simmons. This interpretation of the EU AIA would not make providers of GenAI models that do not comply with EU copyright law liable for copyright infringement. However, they could violate the EU AIA and be subject to the enforcement provisions of that European Union law.
Dr Jani indicated that this interpretation of the EU AIA could raise compliance with international trade agreements by creating a prohibited barrier to trade. However, the Fordham panel did not go any deeper on this issue.