• hperrin@lemmy.ca
    link
    fedilink
    English
    arrow-up
    70
    ·
    1 day ago

    Good news! It is! It’s actually public domain. If you get ahold of some AI generated code, you can do whatever you want with it.

    • bss03@infosec.pub
      link
      fedilink
      English
      arrow-up
      2
      ·
      4 hours ago

      Not really. It doesn’t necessarily get a new copyright, but can be a derivative work of the training data.

      “The training process [also] implicates the right of reproduction.” – Part 3, III, B.

      “outputs likely infringe the reproduction right and, to the extent they adapt the originals, the right to prepare derivative works. Some commenters noted that, depending on the content type and the audience, they may implicate the public display and public performance rights as well.” — Part 3, III, D

      “The Office concludes that, given current generally available technology, prompts alone do not provide sufficient human control to make users of an AI system the authors of the output.” – Part 2, II, D, 2

      “Affirming the Office’s refusal, the court stated that ‘copyright law protects only works of human creation’” – Part 2, II, B

      “In most cases, however, humans will be involved in the creation process, and the work will be copyrightable to the extent that their contributions qualify as authorship.” – Part 2, II, B

      All quotes from: https://www.copyright.gov/ai/

    • NakedNateRollerSkate@lemmy.dbzer0.com
      link
      fedilink
      arrow-up
      49
      ·
      edit-2
      1 day ago

      I’m a lawyer that specializes in this area of IP law. It is true that code generated by an LLM lacks a human author and is therefore not copyright eligible material. However, there are a few nuances that technically make your statement somewhat incorrect.

      One, the code can still be sold or licensed as personal property. However, once somebody gets their hands on that code, there few things stopping them from re-releasing it for free and entering it into the public domain. Even a contract prohibiting such transfer will likely fail due to a small provision in US copyright law.

      Two, this code can still be protected as a trade secret, so don’t think you can create code at work and just freely distribute it if your employer has asked you to keep it confidential.

      There are a number of other possible nuances, but those are two that I see come up most often in my practice.

      • atzanteol@sh.itjust.works
        link
        fedilink
        English
        arrow-up
        2
        ·
        16 hours ago

        I’m a lawyer that specializes in this area of IP law. It is true that code generated by an LLM lacks a human author and is therefore not copyright eligible material.

        That sounds a bit over simplified? I thought the decisions so far were for cases with minimal human input but so far no decisions in cases where a person guides an llm or has significant input into the output.

        • Natanael@infosec.pub
          link
          fedilink
          arrow-up
          1
          ·
          3 hours ago

          The issue here is that the output must contain human expression to be protected. If you ask it to just solve a problem for you then you haven’t put expression in because that’s just functional use. And if it doesn’t respond in kind by preserving your expression then likewise there’s nothing protected.

          On the other hand, if your start with clearly copyright protected works of your own like books or poem collections, then if you ask it to expand on it or edit then you’ll have much more of your actual expression in the output.

          Or, if you simply only use the LLM output as a template and then edit yourself to make it yours.

          • atzanteol@sh.itjust.works
            link
            fedilink
            English
            arrow-up
            1
            ·
            1 hour ago

            The issue here is that the output must contain human expression to be protected. If you ask it to just solve a problem for you then you haven’t put expression in because that’s just functional use. And if it doesn’t respond in kind by preserving your expression then likewise there’s nothing protected.

            Wasn’t that the issue with photography as well? “I just push a button” and the machine takes the photograph. Did I create that photo or did the camera? The expression was deemed to be that of the photographer in this case. But it’s the machine doing the actual work.

            On the other hand, if your start with clearly copyright protected works of your own like books or poem collections, then if you ask it to expand on it or edit then you’ll have much more of your actual expression in the output. Or, if you simply only use the LLM output as a template and then edit yourself to make it yours.

            Right. And what if you guide the LLM through writing the code? e.g. “No not like that - do this instead” - or if you go beyond telling it what you want and tell it how to go about doing it? What if you have style guidelines that you want it to follow? How much of the code has your “expression” in it as a result? Frankly that can be a lot more work than “push a button and the machine does the work”. But whether that work is considered to be the “expression” of the developer or not… I dunno.

            I feel like there’s a lot of gray area here. Everything from “just fully vibe-coded” to “coded along-side an LLM”.

        • NakedNateRollerSkate@lemmy.dbzer0.com
          link
          fedilink
          arrow-up
          5
          ·
          7 hours ago

          The biggest case in this field about human authorship is the Naruto monkey selfie case, but for the cases youre talking about (human input and an LLM) there havr been a number of administrative law cases at the US Copyright office. The two most useful to read are Zarya of the Dawn, Thaler, and Suryast.

      • NaibofTabr@infosec.pub
        link
        fedilink
        English
        arrow-up
        13
        ·
        1 day ago

        OK, but shouldn’t any code descended from GNU GPL code also be GNU GPL? Therefore if a model was trained on any GPL code, any output that it generates should also be GPL.

        • NakedNateRollerSkate@lemmy.dbzer0.com
          link
          fedilink
          arrow-up
          11
          ·
          24 hours ago

          It depends how you use the GPL code. Merely training an LLM with such code is arguably fair use. I personally disagree with that assessment, buy most courts have ruled it as fair use so far. To be honest, US copyright law just merely wasn’t written with modern LLMs in mind and it is desperately in need of a number of statutory updates.

          • NaibofTabr@infosec.pub
            link
            fedilink
            English
            arrow-up
            5
            ·
            23 hours ago

            The training, sure, that’s not what I’m talking about. I’m saying that any code output by a model trained on GPL code must be GPL, because it is derived from GPL code.

            • NakedNateRollerSkate@lemmy.dbzer0.com
              link
              fedilink
              arrow-up
              11
              ·
              23 hours ago

              There is a little bit of confusion here I think. Copyright protects the expression itself (I.e., the code as it is written) and it allows the author (or owner) to stop people from infringing their code, which means performing, displaying, etc… The license – in your example, a GPL license – is just a contract that gives other people permission to do those things without committing copyright infringement. Fair use is a a defense someone can raise when they are being sued for copyright infringement.

              The lawsuits against LLM companies are primarily based on the theory that LLM companies are infringing the copyright of the original authors beyond the scope of any license they might have, if any. The AI companies are saying that their scraping and use of the underlying work to train the model is fair use.

              The thing you are talking about, the code being outputted by the LLMs, is sufficiently disconnected from the original copyrightable expression such that courts have so far ruled that any such output is not a derivative work of the original. This is mainly because such outputs have a randomness/temperature in how they are created, so it isn’t producing the same expression nor is it a derivative work under existing law.

              There are a lot of other factors that go into all of this. For example, under Google v. Oracle, code generally gets “thinner” protection than literary or more creative works because it is functional in nature. There is an entire career’s worth of case law on this topic so it can get quite fact specific. But hopefully my explanation of the current system makes a little more sense.

              • bss03@infosec.pub
                link
                fedilink
                English
                arrow-up
                2
                ·
                4 hours ago

                That doesn’t align with the guidelines published by the U.S. copyright office. They claim outputs similar to training data do infringe on the copyright of the training data, effectively supporting the “derivative work” claim, since they do NOT have to be identical.

                “The training process [also] implicates the right of reproduction.” – Part 3, III, B.

                “outputs likely infringe the reproduction right and, to the extent they adapt the originals, the right to prepare derivative works. Some commenters noted that, depending on the content type and the audience, they may implicate the public display and public performance rights as well.” — Part 3, III, D

                “The Office concludes that, given current generally available technology, prompts alone do not provide sufficient human control to make users of an AI system the authors of the output.” – Part 2, II, D, 2

                “Affirming the Office’s refusal, the court stated that ‘copyright law protects only works of human creation’” – Part 2, II, B

                “In most cases, however, humans will be involved in the creation process, and the work will be copyrightable to the extent that their contributions qualify as authorship.” – Part 2, II, B

                All quotes from: https://www.copyright.gov/ai/

                • Natanael@infosec.pub
                  link
                  fedilink
                  arrow-up
                  2
                  ·
                  3 hours ago

                  Especially image models where you can literally just ask for and get images from copyright protected works like art, movies, comics, etc

              • NaibofTabr@infosec.pub
                link
                fedilink
                English
                arrow-up
                4
                ·
                20 hours ago

                Ah, well, thank you for the explanation. It’s not confusion - I disagree with the court decisions you’re describing.

                Randomness or no, a copy-paste machine is still a copy-paste machine. It doesn’t do anything else. LLMs are language search algorithms with fancy optimization routines bolted on, nothing more. Everything else that’s been said about these “artificial intelligence” products is marketing.

                Reaching a different conclusion about the “transformative nature” (or whatever) of LLM output requires understanding less about how they actually work.

                • atzanteol@sh.itjust.works
                  link
                  fedilink
                  English
                  arrow-up
                  1
                  ·
                  edit-2
                  4 hours ago

                  Randomness or no, a copy-paste machine is still a copy-paste machine. It doesn’t do anything else. LLMs are language search algorithms with fancy optimization routines bolted on, nothing more.

                  This sort of extreme reductionism is not helpful. They are significantly different from a “copy-paste” machine. The original training data does not exist in the model itself. It’s very much a “transformative” process.

                  Reaching a different conclusion about the “transformative nature” (or whatever) of LLM output requires understanding less about how they actually work.

                  This sounds like you’re saying “wilful ignorance is better than knowledge” - surely that’s not what you meant though?

                  • NaibofTabr@infosec.pub
                    link
                    fedilink
                    English
                    arrow-up
                    1
                    ·
                    edit-2
                    2 hours ago

                    They are significantly different from a “copy-paste” machine. The original training data does not exist in the model itself. It’s very much a “transformative” process.

                    This is false. There is a pile of research demonstrating that copies of training data can be extracted from trained models, going back at least 5 years. If the model can reproduce a copy of the original, then it must be stored in the neural network in some form. Here are some examples:

                    […] an adversary can perform a training data extraction attack to recover individual training examples by querying the language model.
                    We demonstrate our attack on GPT-2, a language model trained on scrapes of the public Internet, and are able to extract hundreds of verbatim text sequences from the model’s training data. These extracted examples include (public) personally identifiable information (names, phone numbers, and email addresses), IRC conversations, code, and 128-bit UUIDs. Our attack is possible even though each of the above sequences are included in just one document in the training data.
                    We comprehensively evaluate our extraction attack to understand the factors that contribute to its success. Worryingly, we find that larger models are more vulnerable than smaller models.

                    “Extracting Training Data from Large Language Models”, Aug 2021, Nicholas Carlini et. al., https://www.usenix.org/system/files/sec21-carlini-extracting.pdf

                    This is also true for diffusion models used for image recognition and generation:

                    In this work, we show that diffusion models memorize individual images from their training data and emit them at generation time.

                    “Extracting Training Data from Diffusion Models”, Aug 2023, Nicholas Carlini et. al., https://www.usenix.org/system/files/usenixsecurity23-carlini.pdf

                    […] we develop a new divergence attack that causes the model to diverge from its chatbot-style generations and emit training data at a rate 150x higher than when behaving properly. Our methods show practical attacks can recover far more data than previously thought, and reveal that current alignment techniques do not eliminate memorization.

                    “Scalable Extraction of Training Data from (Production) Language Models” 2023, Milad Nasr et. al., https://www.microsoft.com/en-us/research/publication/scalable-extraction-of-training-data-from-production-language-models/

                    We presented yet another method to extract training data from production language models. Our finetuning attack highlights the fact that existing alignment techniques only provide a false sense of privacy and do not solve the underlying memorization issues.

                    “Extracting (Even More) Training Data From Production Language Models”, July 2024, Javier Rando et. al., https://spylab.ai/blog/training-data-extraction/

                    This is not limited to older models, it is current:

                    With a simple two-phase procedure (Section 3), we show that it is possible to extract large amounts of in-copyright text from four production LLMs.
                    […] long-standing, clear, and sound technical facts remain: LLMs memorize portions of their training data (Carlini et al., 2021; Carlini et al., 2023), these memorized data are encoded in the model’s weights (Nasr et al., 2023; Carlini, 2025; Schwarzschild et al., 2024), and, as we show here, it can be feasible to extract large quantities of in-copyright training data from production LLMs.

                    “Extracting books from production language models”, Jan 2026, Ahmed Ahmed et. al., https://arxiv.org/abs/2601.02671

                    You can find proof-of-concept code on Github:

                    https://github.com/brian-lou/Training-Data-Extraction-Attack-on-LLMs

                    https://github.com/google-research/lm-extraction-benchmark

                    This is a security concern for companies handling proprietary information with LLMs, resulting in security companies offering services to respond to data extraction attacks:

                    https://www.nightfall.ai/ai-security-101/training-data-extraction-attacks

                    There is lots of information about this on the internet, all you have to do is a little bit of your own research.


                    Reaching a different conclusion about the “transformative nature” (or whatever) of LLM output requires understanding less about how they actually work.

                    This sounds like you’re saying “wilful ignorance is better than knowledge” - surely that’s not what you meant though?

                    No, what I mean is that believing that models do not contain copies of training data requires a person to be ignorant of the actual research and the reality of how these models function. If you understand the technical details, then you know that the model weights do in fact encode copies of the training data.

                    The output of an LLM is just regurgitation of compressed data. Breaking the data up into smaller pieces (“tokenization”) does not make it any less of a copy.

                    The result of an AI query is basically a scrapbooker’s collage of more-or-less related clippings from random sources, stitched together and presented as if they were a single coherent(ish) statement. LLMs are automated plagiarism systems.

                  • Natanael@infosec.pub
                    link
                    fedilink
                    arrow-up
                    2
                    ·
                    3 hours ago

                    The ideal is that the training data don’t exist on it. The reality is that the training algorithm is a very very weird compression algorithm and sometimes the model contains just derived properties and sometimes you can perform sample recovery and get entire books back out. That which is recoverable with a prompt less complex than the output is inherently already in the model, because that’s simply how entropy works.

        • JackbyDev@programming.dev
          link
          fedilink
          English
          arrow-up
          2
          ·
          21 hours ago

          Models are trained on copyrighted content without licenses (e.g., just default copyright) and it doesn’t mean anything about the output, for better or worse. So why would something less strict (any sort of permission given past the default copyright protections) change that?

          • bss03@infosec.pub
            link
            fedilink
            English
            arrow-up
            1
            ·
            4 hours ago

            That’s not what the U.S. Copyright Office says.

            “The training process [also] implicates the right of reproduction.” – Part 3, III, B.

            “outputs likely infringe the reproduction right and, to the extent they adapt the originals, the right to prepare derivative works. Some commenters noted that, depending on the content type and the audience, they may implicate the public display and public performance rights as well.” — Part 3, III, D

            “The Office concludes that, given current generally available technology, prompts alone do not provide sufficient human control to make users of an AI system the authors of the output.” – Part 2, II, D, 2

            “Affirming the Office’s refusal, the court stated that ‘copyright law protects only works of human creation’” – Part 2, II, B

            “In most cases, however, humans will be involved in the creation process, and the work will be copyrightable to the extent that their contributions qualify as authorship.” – Part 2, II, B

            All quotes from: https://www.copyright.gov/ai/

          • NaibofTabr@infosec.pub
            link
            fedilink
            English
            arrow-up
            1
            ·
            1 day ago

            Kind of hoping that @NakedNateRollerSkate@lemmy.dbzer0.com has a more concrete answer, if they work professionally in this field, something less trite.

      • TehPers@beehaw.org
        link
        fedilink
        English
        arrow-up
        1
        ·
        19 hours ago

        I’m curious what your thoughts are on the lawsuits against companies scraping YouTube videos for violating DMCA §1201. It’s a different type of (potential) copyright infringement that, from what I understand, ignores the license of the original copyrighted material and focuses on the distribution.

        (Also, for the sake of setting the record, we can assume none of this is legal advice, right?)

      • copygirl@lemmy.blahaj.zone
        link
        fedilink
        English
        arrow-up
        3
        ·
        1 day ago

        And from what I understand, LLMs are known to reproduce some text verbatim. That original text was made by a real human being, and thus would still be protected under copyright.