4 ms·

This is a tired cliche. Does Supabase owe him royalties when they use that same model that trained on Eminem’s lyrics to make a chat bot that can interact with

This is a tired cliche. Does Supabase owe him royalties when they use that same model that trained on Eminem’s lyrics to make a chat bot that can interact with their support documentation? https://supabase.com/blog/chatgpt-supabase-docs https://supabase.com/blog/chatgpt-supabase-docs

Quarrelsome 4y ago
Chat bots are not in competition with Eminem in his industry. Whereas other music artists are. Its entirely fair game for subjects in a given field to protest when an AI is trained on their work and then is used to power competition to them without any form of remuneration or licensing to compensate them.
- williamcotton 4y ago
  Great, so we agree that it is fair use for OpenAI to train on copyrighted works as long as the person using the model does so in non-infringing ways and that the liability rests with the person using the tool. Logically, the model must be considered in fair use if we are to allow Supabase Clippy. This is how the courts will see it!
tsimionescu 4y ago
Possibly yes, though it's more likely that OpenAI owes royalties on the revenue they get from ChatGPT. We have yet to see this play out in court, but we likely will in the following years.
- williamcotton 4y ago
  Universal tried to make the same argument with Sony and the VCR. They wanted a royalty payment on every sale. They won their case! It took the Supreme Court to overrule, where they created the legal doctrine of “commercially significant non-infringing use” as part of determining fair use. They looked at the VHS rental market and the numerous other industries that were dependent on this technology. The courts do not want to be in a position where a ruling can have a profound impact on commerce especially when balanced with extending copyright claims at the detriment to freedoms of speech. This is why legal experts keep saying the model and the outputs will be considered different and why the model itself is fair use and why the risk-averse lawyers at Microsoft are getting behind these tools… and even why they are letting people build on top of them before they were ready for prime time, to flood the market with non-infringing uses!
  - simiones 4y ago
    While I think the case you're presenting is very likely, I don't think Sony v Universal applies too significantly. The VCR was demonstrably not in any way based on Universal's IP - it was only a tool that could be used to create copies of Universal's IP. In their capacity as tools, LLMs probably would fall under a similar model. However, the LLM itself is substantially based on the IP of all of the creators of the data used in its training set. This may or may not be granted fair use status, or it may or may not even be seen as infringing on the copyright of those works at all - but either way, it is not covered by that case as a legal question (at least in my own non-lawyer interpretation).
    - williamcotton 4y ago
      Sony v Universal established a very important legal doctrine with regards to "commercially significant non-infringing use". You can take my word for it, you can go an do your own research, you can confirm with an IP lawyer, or you can wait for the the court's opinion. Or I guess you can give me a little bit of time to go and help you do some of your own research, which I will do right now, so just hold on a bit! I'm not trying to trick people or win some hypothetical argument, I'm trying to help people see how the courts will consider these issues and how and why we should agree with their rulings that these tools are fair use of copyright protected works! I love copyright, probably more than most people on these forums, but the liability needs to be on the people using these tools. Supabase Clippy is fantastic and they should not bear any costs, even implicitly through OpenAI paying royalties. Someone who releases a "Sarah Anderson Cartoon Maker" tool using Stable Diffusion should still be found to wrong Sarah Anderson's common law right to publicity just as someone who releases a "Vacation Photo Background Cleaner-Upper" tool using Stable Diffusion should not bear any costs, even implicitly through StabilityAI paying royalties to Sarah Anderson. These must be considered in their capacities as tools regardless of how they were made and for reasons foundational to the law itself: How can you prove that I used a given tool such as Stable Diffusion or the "Vacation Photo Background Cleaner-Upper" tool without concrete evidence of the tool, such as the presence of the software on my laptop? Can law enforcement get a warrant to search based on literally no visual evidence that Sarah Anderson's works were somehow used in the training process of a tool that I have on my private property? Edit: --- In the Texas Law Review in March, 2021, Mark Lemley, a Stanford law professor, and Bryan Casey, then a lecturer in law at Stanford, posed a question: "Will copyright law allow robots to learn?" They argue that, at least in the United States, it should. "[Machine learning] systems should generally be able to use databases for training, whether or not the contents of that database are copyrighted," they wrote, adding that copyright law isn't the right tool to regulate abuses. But when it comes to the output of these models – the code suggestions automatically made by the likes of Copilot – the potential for the copyright claim proposed by Butterick looks stronger. "I actually think there's a decent chance there is a good copyright claim," said Tyler Ochoa, a professor in the law department at Santa Clara University in California, in a phone interview with The Register. In terms of the ingestion of publicly accessible code, Ochoa said, there may be software license violations but that's probably protected by fair use. While there hasn't been a lot of litigation about that, a number of scholars have taken that position and he said he's inclined to agree. https://www.theregister.com/2022/10/19/github_copilot_copyright/ https://www.theregister.com/2022/10/19/github_copilot_copyri... https://texaslawreview.org/fair-learning/ https://texaslawreview.org/fair-learning/ Both Lemley and Ochoa state that the models themselves are probably protected by fair use. Meaning, it is perfectly fine for OpenAI to train their models on publicly accessible copyright protected works without asking for permission or having to pay any royalties or having to adhere to any of the terms of the license. They are also free to distribute this tool and to charge for people to use this tool. What Ochoa is saying about there being a good chance of there being a copyright claim is that this tool doesn't absolve the users of the tool from copyright violation. The liability is on the person using the tool regardless of the tool being used. It's not the intent that matters with copyright, it's that you ended up publishing something that looks enough like someone else's picture that twelve people consider it to be not too different than a simple photocopy. Now, this can still be a problem for Copilot because an engineer's company might not want to be injecting a lot of copyright protected code into their products, but for the most part the outputs from Copilot have been non-infringing. That it sometimes produces infringing code does not matter to anyone other than the person using Copilot. Ochoa goes on in detail about what is and isn't covered by copyright with regards to code, which is one of the things that gives me confidence to use Copilot and know that I'm not putting myself at risk: But in terms of where Copilot may be vulnerable to a copyright claim, Ochoa believes LLMs that output source code – more so than models that generate images – are likely to echo training data. That may be problematic for GitHub. "When you're trying to output code, source code, I think you have a very high likelihood that the code that you output is going to look like one or more of the inputs, because the whole point of code is to achieve something functional," he said. "Once something works well, lots of other people are going to repeat it." Ochoa argues the output is likely to be the same as the training data for one of two reasons: "One is there's only one good way to do it. And the other is [you're] copying basically an open source solution. "If there's only one good way to do it, OK, then that's probably not eligible for copyright. But chances are that there's just a lot of code in [the training data] that has used the same open source solution, and that the output is going to look very similar to that. And that's just copying." In other words, the model may suggest code to solve a problem for which there's only really one practical solution, or it's copying from someone's open source that does the same thing. In either case, that's probably because a lot of people have used the same code, and that shows up a lot in the training data, leading to the assistant regurgitating it. So in practice it is pretty easy to tell that Copilot is spitting out purely functional suggestions basically all of the time as there isn't really any other way to wire up a unit test or call a specific API. Ironically, if Copilot gets better at "software architecture" then it starts to cross over into the expressive parts of software that are indeed covered by copyright, meaning these issues of liability become harder to discern to the end user and enough of a problem that GitHub would want to figure out attribution or somehow "clear" the suggestions for the user.
- rapsey 4y ago
  OpenAI will 100% need to start paying royalties. For instance it knows a whole lot of things that are not publicly available and locked away behind science journal paywalls.
  - flangola7 4y ago
    I don't see why that would require them to pay royalties. The training data is also publicly available, https://pile.eleuther.ai/ https://pile.eleuther.ai/
    - williamcotton 4y ago
      From The Pile: An 800GB Dataset of Diverse Text for Language Modeling: https://arxiv.org/pdf/2101.00027.pdf https://arxiv.org/pdf/2101.00027.pdf 7.1 Legality of Content While the machine learning community has begun to discuss the issue of the legality of training models on copyright data, there is little acknowledgment of the fact that the processing and distribution of data owned by others may also be a violation of copyright law. As a step in that direction, we discuss the reasons we believe that our use of copyright data is in compliance with US copyright law. Under pre (1984) (and affirmed in subsequent rulings such as aff (2013); Google (2015)), non-commercial, not-for-profit use of copyright media is preemptively fair use. Additionally, our use is transformative, in the sense that the original form of the data is ineffective for our purposes and our form of the data is ineffective for the purposes of the original documents. Although we use the full text of copyright works, this is not necessarily disqualifying when the full work is necessary (ful, 2003). In our case, the long-term dependencies in natural language require that the full text be used in order to produce the best results (Dai et al., 2019; Rae et al., 2019; Henighan et al., 2020; Liu et al., 2018). Copyright law varies by country, and there may be additional restrictions on some of these works in particular jurisdictions. To enable easier compliance with local laws, the Pile reproduction code is available and can be used to exclude certain components of the Pile which are inappropriate for the user. Unfortunately, we do not have the metadata necessary to determine exactly which texts are copyrighted, and so this can only be undertaken at the component level. Thus, this should be be taken to be a heuristic rather than a precise determination. 1984. Sony corp. of america v. universal city studios, inc. 2003. Kelly v. arriba soft corp. 2013. Righthaven llc v. hoehn.

Quarrelsome 4y ago

Chat bots are not in competition with Eminem in his industry. Whereas other music artists are. Its entirely fair game for subjects in a given field to protest when an AI is trained on their work and then is used to power competition to them without any form of remuneration or licensing to compensate them.

williamcotton 4y ago
Great, so we agree that it is fair use for OpenAI to train on copyrighted works as long as the person using the model does so in non-infringing ways and that the liability rests with the person using the tool. Logically, the model must be considered in fair use if we are to allow Supabase Clippy. This is how the courts will see it!

tsimionescu 4y ago

Possibly yes, though it's more likely that OpenAI owes royalties on the revenue they get from ChatGPT. We have yet to see this play out in court, but we likely will in the following years.

williamcotton 4y ago
Universal tried to make the same argument with Sony and the VCR. They wanted a royalty payment on every sale. They won their case! It took the Supreme Court to overrule, where they created the legal doctrine of “commercially significant non-infringing use” as part of determining fair use. They looked at the VHS rental market and the numerous other industries that were dependent on this technology. The courts do not want to be in a position where a ruling can have a profound impact on commerce especially when balanced with extending copyright claims at the detriment to freedoms of speech. This is why legal experts keep saying the model and the outputs will be considered different and why the model itself is fair use and why the risk-averse lawyers at Microsoft are getting behind these tools… and even why they are letting people build on top of them before they were ready for prime time, to flood the market with non-infringing uses!
- simiones 4y ago
  While I think the case you're presenting is very likely, I don't think Sony v Universal applies too significantly. The VCR was demonstrably not in any way based on Universal's IP - it was only a tool that could be used to create copies of Universal's IP. In their capacity as tools, LLMs probably would fall under a similar model. However, the LLM itself is substantially based on the IP of all of the creators of the data used in its training set. This may or may not be granted fair use status, or it may or may not even be seen as infringing on the copyright of those works at all - but either way, it is not covered by that case as a legal question (at least in my own non-lawyer interpretation).
  - williamcotton 4y ago
    Sony v Universal established a very important legal doctrine with regards to "commercially significant non-infringing use". You can take my word for it, you can go an do your own research, you can confirm with an IP lawyer, or you can wait for the the court's opinion. Or I guess you can give me a little bit of time to go and help you do some of your own research, which I will do right now, so just hold on a bit! I'm not trying to trick people or win some hypothetical argument, I'm trying to help people see how the courts will consider these issues and how and why we should agree with their rulings that these tools are fair use of copyright protected works! I love copyright, probably more than most people on these forums, but the liability needs to be on the people using these tools. Supabase Clippy is fantastic and they should not bear any costs, even implicitly through OpenAI paying royalties. Someone who releases a "Sarah Anderson Cartoon Maker" tool using Stable Diffusion should still be found to wrong Sarah Anderson's common law right to publicity just as someone who releases a "Vacation Photo Background Cleaner-Upper" tool using Stable Diffusion should not bear any costs, even implicitly through StabilityAI paying royalties to Sarah Anderson. These must be considered in their capacities as tools regardless of how they were made and for reasons foundational to the law itself: How can you prove that I used a given tool such as Stable Diffusion or the "Vacation Photo Background Cleaner-Upper" tool without concrete evidence of the tool, such as the presence of the software on my laptop? Can law enforcement get a warrant to search based on literally no visual evidence that Sarah Anderson's works were somehow used in the training process of a tool that I have on my private property? Edit: --- In the Texas Law Review in March, 2021, Mark Lemley, a Stanford law professor, and Bryan Casey, then a lecturer in law at Stanford, posed a question: "Will copyright law allow robots to learn?" They argue that, at least in the United States, it should. "[Machine learning] systems should generally be able to use databases for training, whether or not the contents of that database are copyrighted," they wrote, adding that copyright law isn't the right tool to regulate abuses. But when it comes to the output of these models – the code suggestions automatically made by the likes of Copilot – the potential for the copyright claim proposed by Butterick looks stronger. "I actually think there's a decent chance there is a good copyright claim," said Tyler Ochoa, a professor in the law department at Santa Clara University in California, in a phone interview with The Register. In terms of the ingestion of publicly accessible code, Ochoa said, there may be software license violations but that's probably protected by fair use. While there hasn't been a lot of litigation about that, a number of scholars have taken that position and he said he's inclined to agree. https://www.theregister.com/2022/10/19/github_copilot_copyright/ https://www.theregister.com/2022/10/19/github_copilot_copyri... https://texaslawreview.org/fair-learning/ https://texaslawreview.org/fair-learning/ Both Lemley and Ochoa state that the models themselves are probably protected by fair use. Meaning, it is perfectly fine for OpenAI to train their models on publicly accessible copyright protected works without asking for permission or having to pay any royalties or having to adhere to any of the terms of the license. They are also free to distribute this tool and to charge for people to use this tool. What Ochoa is saying about there being a good chance of there being a copyright claim is that this tool doesn't absolve the users of the tool from copyright violation. The liability is on the person using the tool regardless of the tool being used. It's not the intent that matters with copyright, it's that you ended up publishing something that looks enough like someone else's picture that twelve people consider it to be not too different than a simple photocopy. Now, this can still be a problem for Copilot because an engineer's company might not want to be injecting a lot of copyright protected code into their products, but for the most part the outputs from Copilot have been non-infringing. That it sometimes produces infringing code does not matter to anyone other than the person using Copilot. Ochoa goes on in detail about what is and isn't covered by copyright with regards to code, which is one of the things that gives me confidence to use Copilot and know that I'm not putting myself at risk: But in terms of where Copilot may be vulnerable to a copyright claim, Ochoa believes LLMs that output source code – more so than models that generate images – are likely to echo training data. That may be problematic for GitHub. "When you're trying to output code, source code, I think you have a very high likelihood that the code that you output is going to look like one or more of the inputs, because the whole point of code is to achieve something functional," he said. "Once something works well, lots of other people are going to repeat it." Ochoa argues the output is likely to be the same as the training data for one of two reasons: "One is there's only one good way to do it. And the other is [you're] copying basically an open source solution. "If there's only one good way to do it, OK, then that's probably not eligible for copyright. But chances are that there's just a lot of code in [the training data] that has used the same open source solution, and that the output is going to look very similar to that. And that's just copying." In other words, the model may suggest code to solve a problem for which there's only really one practical solution, or it's copying from someone's open source that does the same thing. In either case, that's probably because a lot of people have used the same code, and that shows up a lot in the training data, leading to the assistant regurgitating it. So in practice it is pretty easy to tell that Copilot is spitting out purely functional suggestions basically all of the time as there isn't really any other way to wire up a unit test or call a specific API. Ironically, if Copilot gets better at "software architecture" then it starts to cross over into the expressive parts of software that are indeed covered by copyright, meaning these issues of liability become harder to discern to the end user and enough of a problem that GitHub would want to figure out attribution or somehow "clear" the suggestions for the user.
rapsey 4y ago
OpenAI will 100% need to start paying royalties. For instance it knows a whole lot of things that are not publicly available and locked away behind science journal paywalls.
- flangola7 4y ago
  I don't see why that would require them to pay royalties. The training data is also publicly available, https://pile.eleuther.ai/ https://pile.eleuther.ai/
  - williamcotton 4y ago
    From The Pile: An 800GB Dataset of Diverse Text for Language Modeling: https://arxiv.org/pdf/2101.00027.pdf https://arxiv.org/pdf/2101.00027.pdf 7.1 Legality of Content While the machine learning community has begun to discuss the issue of the legality of training models on copyright data, there is little acknowledgment of the fact that the processing and distribution of data owned by others may also be a violation of copyright law. As a step in that direction, we discuss the reasons we believe that our use of copyright data is in compliance with US copyright law. Under pre (1984) (and affirmed in subsequent rulings such as aff (2013); Google (2015)), non-commercial, not-for-profit use of copyright media is preemptively fair use. Additionally, our use is transformative, in the sense that the original form of the data is ineffective for our purposes and our form of the data is ineffective for the purposes of the original documents. Although we use the full text of copyright works, this is not necessarily disqualifying when the full work is necessary (ful, 2003). In our case, the long-term dependencies in natural language require that the full text be used in order to produce the best results (Dai et al., 2019; Rae et al., 2019; Henighan et al., 2020; Liu et al., 2018). Copyright law varies by country, and there may be additional restrictions on some of these works in particular jurisdictions. To enable easier compliance with local laws, the Pile reproduction code is available and can be used to exclude certain components of the Pile which are inappropriate for the user. Unfortunately, we do not have the metadata necessary to determine exactly which texts are copyrighted, and so this can only be undertaken at the component level. Thus, this should be be taken to be a heuristic rather than a precise determination. 1984. Sony corp. of america v. universal city studios, inc. 2003. Kelly v. arriba soft corp. 2013. Righthaven llc v. hoehn.