A recent article in *Nature* proposes amending copyright licenses as a method to address the misuse of artificial intelligence and reassert human control over creative works. The proposal aims to limit the unchecked use of copyrighted material in the training of AI models.
According to the article, published online on August 25, 2026, current copyright licenses often lack specific provisions regarding AI training. This allows AI developers to utilize vast amounts of copyrighted data without explicit permission or compensation to the rights holders. The article highlights a growing concern that this practice undermines the creative process and potentially infringes on copyright.
The proposed solution involves adding clauses to copyright licenses that clearly define acceptable and unacceptable uses of copyrighted material in the context of AI. These clauses could, for example, require AI developers to obtain licenses for training AI models on copyrighted data, or to share revenue generated from AI-created works with the original copyright holders. The article suggests that such amendments would not only protect the rights of creators but also encourage responsible AI development.
The *Nature* article, identified by DOI 10.1038/d41586-026-02633-y, argues that proactively addressing these issues through copyright law is crucial to ensuring that AI serves as a tool to augment, rather than replace, human creativity.
Read the original coverage
💬 Comments
📜 Comment Policy