In early March, a broad majority of the European Parliament voted in favor of a series of recommendations aimed at ending the current 'wild west phase' of AI training. Until now, tech companies have been training their generative models on a large scale with copyrighted material, often without compensating the original creators. The parliament now wants to find a workable balance between technological innovation and the protection of intellectual property.
Transparency and retroactive compensation
The report focuses on three concrete proposals. First, a compensation system must be established for the use of copyrighted material. Since this practice has been ongoing for years, the parliament proposes that these compensations should apply retroactively. Second, transparency becomes the new norm: AI companies must disclose which data they have used to train their models (something they are already required to do under the European AI Regulation). Finally, the core idea is to return control to the creators; they must be able to decide for themselves whether their work can be used for AI purposes at all.
Thus, the goal is not so much to prevent AI training with copyrighted material; the report primarily aims to establish a system in which rights holders can decide whether their work is used for AI training and receive compensation for it.
Between applause and legal doubts
While organizations such as the European Writers Council welcome the report, there is criticism from the tech sector as well as from (legal) think tanks. The Open Future Foundation points out that while the report identifies the right problems, the path to a practical solution remains unclear. Additionally, critics warn of a piling up of regulations. The report deviates in some respects from the existing Copyright Directive and the AI Act, which could lead to legal complexity in an already complex field.
What does this mean for the cultural sector?
For the cultural sector, this report is an important signal: the political will to protect creators in the AI era is growing. Although the report still needs to be translated into concrete legislation, it is advisable for the sector to act proactively. An important point of attention in further development is the functioning of so-called 'opt-out' mechanisms. At present, the initiative often lies with the creator to indicate through technical means (such as metadata or robot.txt) that their work may not be used for data mining and AI training.
For advocacy organizations, there is a task to monitor the implementation of these systems to ensure they align well with the daily practices of creators. For individual creators, it is advisable to remain vigilant about the terms of use of platforms. The legal frameworks are evolving, and this report is a first step toward a system where prior consent can become the standard.
More news by topic
