Have strong opinions, but I welcome any civil fact-based discussion.

Alt account: /u/[email protected]

  • 5.9K Posts
  • 1.1K Comments
Joined 1 year ago
cake
Cake day: July 9th, 2023

help-circle





















  • All of them are considered in tandem, not individually.

    Considering that OpenAI is making a commercial profit from developing its ML models

    They are losing money during development (all those GPUs are not free and running them costs a lot of energy), they are making the money after it’s trained. Just factual inaccuracy.

    And being used for commercial purpose is not automatic rejection. Take YouTube, where fair use comes up constantly. Almost all the cases are for commercial purpose, but most qualify under fair use.

    #3 also because the model usually ingests the entire work, not just part of it.

    While they are trained on full works, the used work in the result is different. Probably minimal considering the size of the models. The fact that some courts already ruled that “AI” works can’t be copyrighted gives weight to the argument that it’s a unique work.

    It’s very hard to argue that “AI” generated is different from someone looking at the original and making a copy by hand. And since the latter is allowed, by the same token is the former.