{"id":100146,"date":"2025-09-13T19:22:09","date_gmt":"2025-09-13T18:22:09","guid":{"rendered":"https:\/\/www.ictrechtswijzer.be\/?p=100146"},"modified":"2025-09-13T23:06:39","modified_gmt":"2025-09-13T22:06:39","slug":"when-does-ai-training-violate-the-three-step-test-in-copyright-law","status":"publish","type":"post","link":"https:\/\/www.ictrechtswijzer.be\/en\/wanneer-is-ai-training-in-strijd-met-de-driestappentoets-in-het-auteursrecht\/","title":{"rendered":"When does AI training violate the three-step test in copyright law?"},"content":{"rendered":"

Generative AI systems are \"trained\" by analyzing vast amounts of data. From this data, which often consists of data collected by copyright<\/a> protected material, the AI learns patterns and styles to create its own new, unique content. According to European legislation, such as the DSM Directive<\/a> , this is under certain conditions allowed<\/a> via an exception for Text and Data Mining (TDM). This possibility was introduced in the Belgian Code of Economic Law (CEL)<\/a> by means of two exceptions: one for research organizations for scientific research (Art. XI.191\/1, 7\u00b0CEL) and one for everyone for all other purposes (Art. XI.190, 20\u00b0 CEL). Under this second exception, TDM is allowed for commercial purposes, such as training commercial AI models. However, authors and other rights holders do have an opt-out possibility under this second exception and thus the ability to oppose it. Separately, the question arises whether or not the use of the two TDM exceptions for AI training is consistent with the three-step test?<\/p>\n\n\n\n

The three-step test<\/h2>\n\n\n\n

Any copyright exception<\/a>, including those for TDM, must meet the so-called three-step test. This test is a cornerstone of international copyright law, found in the Berne Convention, the TRIPS Treaty and various EU directives, and is also found in Article XI.192\/3 CEL. The test states that an exception is valid only if three cumulative conditions are met:<\/p>\n\n\n\n

    \n
  1. It should only concern certain, special cases.<\/li>\n\n\n\n
  2. It should not conflict with a normal exploitation of the work.<\/li>\n\n\n\n
  3. It must not unreasonably prejudice the legitimate interests of the author.<\/li>\n<\/ol>\n\n\n\n

    The discussion of AI training focuses mainly on the second step. According to an influential interpretation by an WTO Panel<\/a> in 2000, \"all forms of exploiting a work, which have, or are likely to acquire, considerable economic or practical importance\" must remain reserved to authors. So the question is whether TDM for AI training is such a form of exploitation.<\/p>\n\n\n\n

    Legal analysis and interpretation: a two-pronged approach<\/h2>\n\n\n\n

    To assess the impact of AI training, we must make an essential distinction between the input (training the AI with existing works) and the output (the content newly generated by the AI).<\/p>\n\n\n\n

    The input<\/h3>\n\n\n\n

    At first glance, it seems logical: If AI developers use works, they would have to pay a license for them. A TDM exception deprives authors of this revenue, which would be a violation of normal exploitation.<\/p>\n\n\n\n

    Yet this argument is legally weak for two reasons:<\/p>\n\n\n\n

      \n
    1. The test focuses on the micro-level<\/strong>: The three-step test requires an analysis of the exploitation of \"a work,\" that is, at the level of a single individual creation. In practice, the potential licensing income for the use of a single book or a single photograph for TDM will never be of \"considerable economic importance.\" Substantial revenue arises only when licensing entire portfolios or catalogs, but the test is not at this macro level.<\/li>\n\n\n\n
    2. The opt-out neutralizes the conflict<\/strong>: The ability for rights holders to reserve (opt-out) their rights to AI training for commercial purposes is a legally important one. If an author shields his work, the AI developer may not use it without permission. As a result, there can be no conflict with normal exploitation at the input stage; indeed, the author retains complete control over this potential licensing market. If an author does not set up an opt-out, it can be argued that he or she accepts the risk of insufficient demand for a TDM license.<\/li>\n<\/ol>\n\n\n\n

      The output<\/h3>\n\n\n\n

      The discussion then shifts to output. What if an AI, trained on the work of authors A, B and C, generates content so good that no one wants to buy author A's new work? This undermines the market and thus normal exploitation.<\/p>\n\n\n\n

      The EU Court of Justice has in previous cases (ACI Adam<\/a>, Movie Player<\/a>) held that an exception harms normal exploitation if it encourages the distribution of illegal copies, leading to fewer sales of legal works.<\/p>\n\n\n\n

      However, there is a crucial difference with GenAI:<\/p>\n\n\n\n