GEMA's AI charter: creative, human performance is the basis of AI

Fair conditions and exploiting the potential of generative AI at the same time – These are two of the key points that have prompted GEMA to present an AI charter. The charter is intended as food for thought and a guideline for the responsible use of AI. The Gesellschaft für musikalische Aufführungs- und mechanische Vervielfältigungsrechte, or GEMA for short, cannot issue anything binding or binding. Nevertheless, it should be able to influence regulators.

The charter comprises ten principles. The first is called: "Digital humanism – puts people first". This is based on the view that AI is a tool and can help people in the creative process, but cannot replace them. The "protection of intellectual property" is intended to protect creators from commercial exploitation and misuse of their works. GEMA also expects a "fair share of the added value".

GEMA principles correspond in part to the AI Regulation

"Transparency" is also a principle of the charter. GEMA is convinced that it must be clear which data has been used to train the AI models. And measures must be taken to ensure that this has been done in accordance with applicable law. The GEMA principles provide for a right of inquiry – It should be possible to request training data. There should be licenses to use works. In addition, GEMA says that when people interact with AI, it must be labeled. This is also what the AI Act says. In addition to notices when someone communicates with an AI chatbot, for example, the European AI regulation also requires AI providers to disclose more information about the training data.

"Negotiations on an equal footing" – This principle is about the imbalance between the large AI providers from the USA, who have a lot of money at their disposal, and other market participants, such as GEMA and other individual stakeholders. It calls for "clear guidelines, competition law instruments and, above all, the strengthening of collective negotiations" in order to pool interests.

"Personal rights" should be respected and "cultural diversity" preserved. Another principle states that "companies that offer AI are accessible for European legal requirements". Initially, this is simply European law. However, there is also the requirement that the law is not undermined by the fact that the training takes place in other regions of the world.

AI should be sustainable. GEMA and the AI Regulation are also intertwined here. AI providers should ultimately take responsibility. "These companies must be aware of the impact of their technology and take responsibility for it. They therefore rightly do not enjoy a liability privilege ('safe harbor'). They must not shift their responsibility onto users."

(emw)