(24 Nov 2025) The backlash from researchers whose work has been used to train large language models (LLMs) has re-opened questions around how scholarly works are licensed and the merits of open licenses. Considering the legal standing of open creative commons licenses and conventional copyright, Martin Paul Eve suggests legal protections for academic work are unlikely to be forthcoming.
“for those who want the biggest spread of openly accessible material, it makes no sense to allow large corporate publishers, for example, to produce closed volumes that reincorporate adapted openly accessible material without forcing them to make their own material open in turn. That is where the power of the share-alike license could be a huge asset in the arsenal of those who want to see the spread of open.
As a result of this complexity, at the Knowledge Commons project at Michigan State University, where I work, we have already seen an influx of questions about whether an open-access license can help to prohibit re-use for machine learning. After all, the Creative Commons Attribution Non-Commercial License sounds as though it might be the perfect tool for such a use case. What, people have also been asking, about the Creative Commons Attribution No Derivatives License?”
Read more here.
