https://www.jipitec.eu/jipitec/article/download/415/417/2163 - discusses the question of licensing of the training data surviving the training process.
https://www.gov.uk/government/consultations/artificial-intelligence-and-intellectual-property-call-for-views/artificial-intelligence-call-for-views-copyright-and-related-rights - gives some discussion from the position of the UK government. The particularly interesting is the statement that in the case that copyright law prevents the development of AI then the copyright law and its exceptions should be adjusted. Also the concept that AI internal processes should in the end be considered from the law perspective the same as internal processes of a human head. Including a point that an AI may remember a particular copyrighted text, just like a human can, and outputting that text with sufficient similarity would be a copyright violation in *both* cases. But that does make either the AI nor the human to be a derivative work of the text they read. https://academic.oup.com/jiplp/article/20/3/182/7922541#509162451 has more discussion on the various exceptions in different jurisdictions with a more sceptical tone On Tue, 6 May 2025 at 21:14, Aigars Mahinovs <aigar...@gmail.com> wrote: > > BTW, this also inspired me to read up on the EU Artificial > Intelligence Act ( https://artificialintelligenceact.eu/ ) and I > noticed a very relevant notice in that text - Art 53 ( > https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32024R1689#art_53 > ) talks about providers of AI having to have provisions to comply with > Article 4(3) of Directive (EU) 2019/790 ( > https://eur-lex.europa.eu/eli/dir/2019/790/oj/eng#art_4 ) and that is > a provision that specifies that (in the context of text and data > mining) there should be a way for rights holders to withhold their > content from the datamining, possibly by specifying a machine-readable > flag for content published online. > > But the key here is that this whole article 4 provides a very explicit > exception to copyright protections for the purposes of text and data > mining. And the EU AI Act very explicitly references this exception as > applicable for AI training purposes. > > It is also nice to see that EU AI Act explicitly highlights open > source AI models and provides them with simplified and preferential > rules. > > On Tue, 6 May 2025 at 01:26, Aigars Mahinovs <aigar...@gmail.com> wrote: > > > > This one is much simpler. Maybe because the lawyers being used are not too > > good. > > > > https://www.courtlistener.com/docket/67538258/tremblay-v-openai-inc/ > > > > Authors claim a lot of stuff, basically a generic shotgun of copyright > > claims, but all secondary claims get dismissed by the court at pre-trial > > stage due to bad legal reasoning and failing to detail or prove any actual > > wrongdoing. And specifically a claim that all outputs from a LLM are > > derived works of all inputs is dismissed based on already decided case law. > > > > Only the claim of direct copyright infringement of using a text of a book > > in the training process of a model still stands to avait the actual trial. > > And there OpenAI is citing a lot of good reasons why that does not > > constitute distribution at all and why the result of the work is > > transformative and thus is protected by fair use. Just the fact of > > accessing some data at some point does not create copyright infringement. > > The whole lawsuit is very sloppy IMHO, IANAL. > > > > On Tue, 6 May 2025 at 00:10, Bill Allombert <ballo...@debian.org> wrote: > >> > >> Le Mon, May 05, 2025 at 11:44:30PM +0200, Aigars Mahinovs a écrit : > >> > On Sun, 4 May 2025 at 17:30, Wouter Verhelst <w...@uter.be> wrote: > >> > > >> > > It is incorrect, because the New York Times did in fact file suit > >> > > against Microsoft, OpenAI, and other parties related to copyright > >> > > infringement of their large library of news articles in creating > >> > > ChatGPT[1]. The case is still in court. > >> > > > >> > > [1] > >> > > https://www.courtlistener.com/docket/68117049/the-new-york-times-company-v-microsoft-corporation/ > >> > > >> > > >> > Thanks for this link, it has been a very interesting read. > >> > >> Another one: > >> > >> https://arstechnica.com/information-technology/2023/07/book-authors-sue-openai-and-meta-over-text-used-to-train-ai/ > >> https://arstechnica.com/tech-policy/2025/02/meta-torrented-over-81-7tb-of-pirated-books-to-train-ai-authors-say/ > >> > >> Cheers, > >> Bill. > >> > > > > > > -- > > Best regards, > > Aigars Mahinovs > > > > -- > Best regards, > Aigars Mahinovs -- Best regards, Aigars Mahinovs
