A research paper commissioned by the European Parliament advocates for EU legislation mandating payment to writers, musicians, and artists for the use of their work in training generative AI models. Current laws surrounding text-and-data mining do not adequately cover the nature of generative AI training, which often distorts existing copyright exceptions. Arguments for fair use in this context are deemed insufficient. The research clarifies that AI does not learn like humans and legally processes copyright-protected content beyond permitted exceptions. A proposal for new EU-level standards is included in the findings.
The study finds the current exception for text-and-data mining in EU law "was not designed to accommodate the expressive and synthetic nature of generative AI training."
"Under EU copyright law, the study finds that such a comparison does not hold. When generative AI models are trained on protected content, they typically make copies and process the actual expressions found in those works."
"While it is often suggested that AI systems 'learn' in ways similar to humans, this analogy is misleading from a legal perspective."
AI acts without understanding - it follows statistical patterns rather than engaging with meaning. This difference matters legally."
Collection
[
|
...
]