3 ms·

So the somewhat necessary followup question: if Midjourney or ChatGPT includes a license in front of its model weights or access to its model that forbids using

So the somewhat necessary followup question: if Midjourney or ChatGPT includes a license in front of its model weights or access to its model that forbids using those weights or output from its model to create a competitor, is it fair use to ignore that license? In both cases, we're relying on copyright. The companies are saying that you can access the models under a license. It's not too hard to circumvent that license or get access through a third party or separate service: but doing so would obviously be seen as deliberate circumvention. We can easily compare that to an artist putting up a click-through in front of a gallery that says that by clicking 'agree' you agree not to use their work for training. And in fact, it's literally the same restriction in both cases: OpenAI maintains that they can use copyright to enforce that people accessing their model avoid using their model to train AIs. Paradoxically, they also maintain that artists can not use copyright to block people accessing their images from using those images to train AIs. Facebook has released model weights with licenses that restrict how those model weights are used. But I can get those models without clicking 'agree' on that license agreement. They're mirrored all over the place. We'll see what courts say, but Facebook's public argument is that their license doesn't stop applying if I download their software from a mirror. So I find it hard to believe that Facebook honestly believes in a fair use argument if they also claim that the same fair use argument doesn't apply to their own copyrighted material that they've released online (assuming that model weights can even be copyrighted in the first place). This is one of my biggest issues with the fair use argument -- it's not that there's nothing convincing about it, in a vacuum I'd very likely agree with it. I don't think that licenses should allow completely arbitrary restrictions. But I also can't ignore that the vast majority of companies making this argument are making it extremely inconsistently. They know that their business models literally don't work if competitors figure out ways to use their services to rapidly produce competing models. If another company comes up with a training method to rapidly replicate model weights or to augment existing models using their output, none of these companies will be OK with that. None of these companies believe that fair use invalidates license terms when it comes to their own IP.

andy99 3y ago
I think companies like openAI, meta and midjourney (if they say you can't train on output) are hypocrites, and probably have no standing to enforce some of the restrictions they try and make. That doesn't change fair use. There are also many truly open source models that explicitly give the users freedom to use them as they want, and I'd prefer more providers license their's like that, ie Apache 2.0 or MIT.
- danShumway 3y ago
  Definitely it doesn't change the fair use argument, but it should change the way we talk about these companies and the way we contextualize what they're arguing and the way we talk about the surrounding debate. We should be up-front that there are plenty of examples where people are not granted these fair use rights and that fair use is often unequally applied to IP law. And we should be pushing hard for the entire package (not to say that you're not, more just pointing out that this is a package deal) -- I am not a copyright maximalist, but I especially don't want a model where only artists and individuals are subject to copyright restrictions. I'm of the opinion that it is good for these companies to take a credibility hit for their hypocrisy and for us to make it clear that the fair use argument has to be something that applies to everybody. I hope that this hypocrisy is something that comes up in court if these lawsuits go through, and ideally I'd like it to become a sticking point where trying to have their cake and eat it too does hurt these companies' legal arguments. I want limited copyright across the board, but a system with maximum fair use for companies like OpenAI and minimal fair use for consumers and individuals is worse than a system with expansive copyright for everyone. I think it's a good thing for companies like OpenAI to be unable to make selective arguments. We shouldn't treat their end goals as irrelevant. It's also very true that there are open models that take a permissive approach, but those models are in my experience usually not the primary target of most criticism (outside of the "open models are dangerous" FUD). The linked comic explicitly calls out "profit-making companies that own the most-used channels of communication." I'm not sure the artist would phrase it this way, but I think reading between the lines the artist is very likely coupling together ownership with their complaint -- they are talking about the injustice of having their work appropriated into a form that is locked down and controlled and that exists for the benefit of the profit-making companies who want to own yet another channel of communication (in this case artistic communication itself). I suspect that part of that feeling of injustice does directly come from an understanding that as an individual anyone would be sued into oblivion if they tried to make the same arguments in the opposite direction. Does that get rid of the fair use argument? No. But it is impossible to actually talk about the fair use debate in a useful way while ignoring the context of that debate. To just mention fair use and nothing else is leaving out a ton of context about why artists are mad, and about why this feels unjust to so many people, and about the regular inconsistencies in how these arguments are often applied. We don't have to get rid of fair use arguments, but I personally think it's more responsible to include that context. I worry that leaving that context out is playing directly into the hands of companies that are trying to create an outcome of selective enforcement for IP. ---- It's a bit of a sidenote, but on the subject of truly open models, I will say that it is great that those models are licensed the way that they're licensed, but I'd still rather have legal recognition that algorithmically generated weights are not subject to copyright at all. It's nice when companies are permissive, but it also subtly implies that they have a choice whether to be permissive. As someone who favors expanding fair use, I would prefer that they not have that choice. I disagree with the application of any license to model weights, permissive or not. I think at best those model weights could be subject to a database license, maybe, but I'm honestly a little skeptical of even that: I think that model weights are a factual and reproducible artifact of training, not a creative work.

andy99 3y ago

I think companies like openAI, meta and midjourney (if they say you can't train on output) are hypocrites, and probably have no standing to enforce some of the restrictions they try and make. That doesn't change fair use. There are also many truly open source models that explicitly give the users freedom to use them as they want, and I'd prefer more providers license their's like that, ie Apache 2.0 or MIT.

danShumway 3y ago
Definitely it doesn't change the fair use argument, but it should change the way we talk about these companies and the way we contextualize what they're arguing and the way we talk about the surrounding debate. We should be up-front that there are plenty of examples where people are not granted these fair use rights and that fair use is often unequally applied to IP law. And we should be pushing hard for the entire package (not to say that you're not, more just pointing out that this is a package deal) -- I am not a copyright maximalist, but I especially don't want a model where only artists and individuals are subject to copyright restrictions. I'm of the opinion that it is good for these companies to take a credibility hit for their hypocrisy and for us to make it clear that the fair use argument has to be something that applies to everybody. I hope that this hypocrisy is something that comes up in court if these lawsuits go through, and ideally I'd like it to become a sticking point where trying to have their cake and eat it too does hurt these companies' legal arguments. I want limited copyright across the board, but a system with maximum fair use for companies like OpenAI and minimal fair use for consumers and individuals is worse than a system with expansive copyright for everyone. I think it's a good thing for companies like OpenAI to be unable to make selective arguments. We shouldn't treat their end goals as irrelevant. It's also very true that there are open models that take a permissive approach, but those models are in my experience usually not the primary target of most criticism (outside of the "open models are dangerous" FUD). The linked comic explicitly calls out "profit-making companies that own the most-used channels of communication." I'm not sure the artist would phrase it this way, but I think reading between the lines the artist is very likely coupling together ownership with their complaint -- they are talking about the injustice of having their work appropriated into a form that is locked down and controlled and that exists for the benefit of the profit-making companies who want to own yet another channel of communication (in this case artistic communication itself). I suspect that part of that feeling of injustice does directly come from an understanding that as an individual anyone would be sued into oblivion if they tried to make the same arguments in the opposite direction. Does that get rid of the fair use argument? No. But it is impossible to actually talk about the fair use debate in a useful way while ignoring the context of that debate. To just mention fair use and nothing else is leaving out a ton of context about why artists are mad, and about why this feels unjust to so many people, and about the regular inconsistencies in how these arguments are often applied. We don't have to get rid of fair use arguments, but I personally think it's more responsible to include that context. I worry that leaving that context out is playing directly into the hands of companies that are trying to create an outcome of selective enforcement for IP. ---- It's a bit of a sidenote, but on the subject of truly open models, I will say that it is great that those models are licensed the way that they're licensed, but I'd still rather have legal recognition that algorithmically generated weights are not subject to copyright at all. It's nice when companies are permissive, but it also subtly implies that they have a choice whether to be permissive. As someone who favors expanding fair use, I would prefer that they not have that choice. I disagree with the application of any license to model weights, permissive or not. I think at best those model weights could be subject to a database license, maybe, but I'm honestly a little skeptical of even that: I think that model weights are a factual and reproducible artifact of training, not a creative work.