A new research paper laid out ways in which AI developers should try and avoid showing LLMs have been trained on copyrighted material.

OpenAI now tries to hide that ChatGPT was trained on copyrighted books, including J.K. Rowling’s Harry Potter series::A new research paper laid out ways in which AI developers should try and avoid showing LLMs have been trained on copyrighted material.

  • uis@lemmy.worldEnglish
    2·
    10 months ago

    Except corporations have infinitely more resources(money, lawyers) compared to people who create. Take Jarek Duda(mathematician from Poland) and Microsoft as an example. He created new compression algorythm, and Microsoft came few years later and patented it in Britain AFAIK. To file patent contest and prior art he needs 100k£.

    • assassin_aragorn@lemmy.worldEnglish
      1·
      10 months ago

      I think there’s an important distinction to make here between patents and copyright. Patents are the issue with corporations, and I couldn’t care less if AI consumed all that.

      • uis@lemmy.worldEnglish
        2·
        10 months ago

        And for copyright there is no possible way to contest it. Also when copyright expires there is no guarantee it will be accessable by humanity. Patents are bad, copyright even worse.