You probably need some technical restrictions as well, but from the legal perspective: is there a license that is like Creative Commons EXCEPT for use cases like use the content for training an LLM by OpenAI or google?

Cc @nostupidquestions@lemmy.world

  • nothacking@discuss.tchncs.de
    21·
    1 year ago

    Would CC BY-SA do? It requires attribution, and there is no real way to attribute text from an LLM to the training data. Assuming that LLM output counts as a derivative work of the training data, which a consensus on has not be reached. (It would also be required for any other copyright based solutions)

    Edit: reading the licence, it requires “a credit identifying the use of the Work in the Derivative Work”, which may allow companies to just post a huge page showing all the training input, but I doubt this counts as “reasonable means”.

    • j4k3@lemmy.worldEnglish
      51·
      1 year ago

      I don’t think anything will get restricted as far as AI inputs for data. The base data doesn’t really exist any more than the collective social consciousness of any given work, artist, or author. Restricting the way AI data is collected will lead to ownership of styles and adjacent works on a level that is extremely draconian. If you post anything publicly in any way, it is essentially a license for AI to use it in the same way it is a license for people to be aware of its existence…IMO…as someone actually playing with offline AI stuff daily right now and noting its limitations. This is era changing tech that is never going away. It’s far better to learn to use it. All the commercial AI stuff is dying anyways. The open source side has already won, the memo is just in transit. LLMs are the framework, not the product.

    • rufus@discuss.tchncs.de
      3·
      1 year ago

      I’d be fine if the ‘SA’ part forces them to share their model with me. They may scrape my website if I get the model in turn.