Breakthrough in copyright: AI data training now legally secured!

Breakthrough in copyright: AI data training now legally secured!

San Francisco, USA - A very hot topic are the current ongoing legal proceedings around the use of copyrighted materials for training artificial intelligence (AI). In the United States, the companies have recently gained the upper hand in an exciting judgment. Their procedures that were directed against various copyright lawsuits are likely to have a significant impact on the entire AI industry. It was specifically about whether the use of copyright -protected books without permission is legal or not. While the judgments of district judges in the United States are still pioneering, there are still numerous lawsuits against sizes such as Google and Openai. reports that ...

District judge William Alsup decided in the Anthropic case that the use of the books was classified as transformative, which is an advantage for AI development. This is opposed to the judgment of Vince Chhaubria for Meta, which also posed positive for the company, but made it clear that the plaintiffs could not provide enough evidence. Here it becomes clear that the judgments support different interpretations of the fair use and only apply to the specific procedures, but not for the general use by the companies themselves.

legal situation in Germany

But not only in the USA is talking about the legal framework for AI. In September 2024, the Hamburg Regional Court decided that research organizations were allowed to use copyright-protected works for AI training under certain conditions. At its core, this was about a photographer who saw his pictures used without consent. However, the research organization referred to various barrier regulations that are used for scientific research.

In the decision of the LG Hamburg, it was clarified that the fleeting reproduction does not represent copyright violation, since the data was deliberately deleted. This strengthens the freedom of science and a balance between copyrights and the need for innovation. The copyright provisions in Germany thus offer a legal framework that gives research organizations more freedom in AI training.

the global perspective

The discussion about this topic runs across the world, and the differences in the respective legal systems are enormous. While in the United States, progress is clearly made, shaped by judgments such as the Bartz v. Anthropic, where the training on protected works has been regarded as a transformative use, the legal discourse does not end there. Techzeitgeist analyzes that …

The global AI market will be estimated at over 100 billion EUR by the end of 2025. However, this enormous market dynamics are in the area of ​​tension between the need for a robust license management and the growing concerns of authors and publishers that fear losing their creative work and not being appropriately remunerated.

Another aspect is the rather reserved regulation in the EU, which is characterized by the AI ​​Act and specific transparency rules. Here the speed of innovation in the AI ​​sector could stagnate, while countries such as China and Japan have other, less strict approaches.
These legal rights will determine the setting for AI and rights in the coming years.

Overall, it should be noted that the disputes about the use of copyrighted materials for AI training are far from complete. The current legal situation can be seen as a first step towards more clarity and legal certainty to promote innovation without endangering creativity. It remains to be seen how the different legal framework will develop and which new markets for licenses and training data could arise.