Landmark ruling on AI training data – Hamburg Regional Court sets standards

On September 27, 2024, the Hamburg Regional Court issued a groundbreaking ruling (case no. 310 O 227/23) in the area of AI training data. This ruling could have far-reaching implications for the use of copyrighted works for the training of AI systems.

Key points of the ruling

1 The court dismissed the action brought by a photographer against the LAION e.V. association.
2. it was decided that the use of copyrighted images for training AI systems without the consent of the rights holders may be permissible under certain circumstances.
3. the court considered the use of the images to train the AI to be covered by the limitation provision of Section 60d UrhG (text and data mining for scientific research).

Reason

Applicability of the research barrier:
The court affirmed the application of the research barrier (Section 60d UrhG). It argued that the creation of a data set for AI training can be considered scientific research as long as it is aimed at gaining knowledge at a later date. The court stated:

Scientific research generally refers to the methodical-systematic pursuit of new knowledge […] The term scientific research, by allowing the methodical-systematic “pursuit” of new knowledge to suffice, is not to be understood so narrowly that it would only cover the work steps directly associated with the acquisition of knowledge; rather, it is sufficient that the work step in question is aimed at a (later) acquisition of knowledge, as is the case, for example, with numerous data collections that must first be carried out in order to subsequently draw empirical conclusions.

Non-commercial purpose:
The court considered the defendant’s activity to be non-commercial because the data set created was made publicly available free of charge.

No decisive influence by companies
The court denied a decisive influence of private companies on the defendant that would exclude the application of the research barrier.

Significance of the judgment

1. precedent: This is one of the first judgments in Germany to deal specifically with the use of copyrighted works for AI training.

2. strengthening AI research: The ruling strengthens the position of research organizations and promotes the development of AI technologies.

3. interpretation of the research barrier: It provides an important interpretation of the research barrier in the context of AI training.

Other important aspects of the judgment

The court also dealt intensively with the question of whether the limitation provision of Section 44b UrhG (text and data mining) was applicable. Although this was not relevant to the decision in the specific case, the court provided important information on the interpretation of this provision

Finally, insofar as it is argued for a teleological reduction of the limitation provision of Section 44b UrhG that the European legislator “simply did not yet have the AI problem” “on its radar” in 2019 when the underlying directive provision (Art. 4 DSM Directive) was created […], this finding alone is clearly not sufficient for a teleological reduction.

The court also emphasized the importance of technological development for the interpretation of the term “machine readability” in the context of reservations of use:

In the Chamber’s view, it would also be a certain contradiction of values to allow the providers of AI models to develop increasingly powerful text-understanding and text-creating AI models via the barrier in Section 44b (2) UrhG on the one hand, but not to require them to use existing AI models within the framework of the barrier in Section 44b (3) sentence 2 UrhG on the other.

Conclusion and outlook

This ruling by Hamburg Regional Court sets an important precedent in the legal assessment of AI training data. It shows that the development of AI technologies can be possible under certain circumstances without the explicit consent of the authors, as long as it is carried out in the context of scientific research.

It is expected that in the future:

1. research organizations could have more legal certainty in the use of training data.
2. there will continue to be discussions about the balance between technological progress and the protection of intellectual property.
3. there may be further legal clarifications or legal adjustments to meet the specific challenges of AI technology in copyright law.

However, the ruling underlines the need for a differentiated view of the use of copyrighted works in the context of AI development. It remains to be seen whether the decision will be upheld in higher instances and how it will influence the further development of the AI sector and copyright law.