Comment by efitz
3 years ago
This is the DRM problem again.
However much we might wish that it was not true, ideas are not rivalrous. If you share an idea with another person, they now have that idea too.
If you share words on paper, then someone with eyes and a brain might memorize them (or much more likely, just grasp and retain the ideas conveyed in the words).
If you let someone hear your music, then the ideas (phrasing, style, melody, etc) in that music are transferred.
If you let people see a visual work, then the stylistic and content elements of that work are potentially absorbed by the audience.
We have copyright to protect specific embodiments, but mostly if you try to share ideas with others without letting them use the ideas you shared, then you are in for a life of frustration and escalating arms race.
I completely sympathize with anyone who had a great idea and spent a lot of effort to realize it. If I invented/created something awesome I would be hurt and angry if someone “copied” it. But the hard cold reality is that you cannot “own” an idea.
> But the hard cold reality is that you cannot “own” an idea.
The above comment is true about the properties of information, as explained via the lens of economics. [1]
However, one ignores ownership as defined by various systems (including the rule of law and social conventions) at one's own peril. Such systems can also present a "hard cold reality" that can bankrupt or ostracize you.
[1] Don't let the apparent confidence and technicality of the language of economists fool you. Economics isn't the only game in town. There are other ways to model and frame the world.
[2] Dangling footnote warning. I think it is instructive to recognize that the field of economics has historically shown a kind of inferiority complex w.r.t. physics. Some economists ascribe to the level of rigor found in physics and that is well and good, but perhaps that effort should not be taken too seriously nor too far, since economics as a field operates at a different level. IMO, it would be wise for more in the field to eat a slice of humble pie.
[3] Ibid. It is well-known that economists can be "hired guns" used to "prove" a wide variety of things, many of which are subjective. My point: you can hire an economist to shore up one's political proposals. Is the same true of physicists? Hopefully not to the same degree. Perhaps there are some cases of hucksterism, but nothing like the history of economists-wagging-the-dog! At some point, the electron tunnels or it does not.
There are other games in town.
But whatever game gives the most predictive power is going to win.
By that metric, various economic schools have been hilariously inept and would get classified not dissimilar to various schools of religious theology with their own dogmas. It's only in the last 15 years or so that some focus on empiricism and explaining reality rather than building theoretical castles in the air is coming about and is still far from mainstream.
There is no need to frame this as "winning versus losing" regarding the many models that we draw upon.
Even when talking about various kinds of scientific and engineering fields, predictive power isn't the only criteria, much less the best. Sometimes the simpler, less accurate models work well enough with less informational and computational cost.
Even if we focus on prediction (as opposed to say statistical inference), often people want some kind of hybrid. Perhaps a blend of satisficing with limited information, scoped action spaces, and bounded computation; i.e. good enough given the information we have to make the decisions we can actuate with some computational budget.
Many terms of art from economics are probably not widely-known here.
> In economics, a good is said to be rivalrous or a rival if its consumption by one consumer prevents simultaneous consumption by other consumers, or if consumption by one party reduces the ability of another party to consume it. - Wikipedia: Rivalry (economics)
Also: we should recognize that stating something as rivalrous or not is descriptive (what exists) not normative (what should be).
I think ideas being rivalrous is intrinsic, and therefore descriptive and normative.
I'm either not understanding you or disagreeing. You seem to be saying that something should be because it is? Saying that would be rather silly, as in "Electrons should repel each other because they repel each other." Not to mention that this claim runs amok of the naturalistic fallacy. So what are you driving at?
2 replies →
We're not trying to keep the AI from learning general ideas, we're trying to keep it from memorizing specific expressions[0]. There's a growing body of research to show that these models are doing a lot of memorizing, even if they're not regurgitating that data. For example, Google's little "ask GPT to repeat a word forever" trick, which will make GPT-4 spit out verbatim training data[1].
If there was a training process that let us pick a minimal sample of examples and turn it into a general purpose art generator or text generator, I think people would have been fine with that. But that's not what any of these models do. They were trained on shittons of creative expression, and there's statistical evidence that the models retain that expression, in a way that is fundamentally different from how humans remember, misremember, adapt, remix, and/or "play around with" other people's creativity.
[0] You called these "embodiments", but I believe you're trying to invoke the idea/expression divide, so I'll run with that.
[1] Or at least it did. OpenAI now filters out conversations that trip the bug.
I don't see the parallel between this offensive tool and DRM. I could, say buy a perpetual license to an image from the artist, so that I can print it and put it on my wall, while it can simultaneously be poisonous to an AI system. I can even steal it and print it, while it is still poisonous to an AI system.
The closest parallel I can think of is that humans can ingest chocolate but dogs should not.
What you've described is the literal, dictionary definition of Digital Rights Management - a technology to restrict the use of a digital asset beyond the contractually-agreed terms. Copying is only one of many uses that the copyright-holder may wish to prevent. The regional lockout on a DVD had nothing to do with copy-protection, but it was still DRM.
It's about the arm's race: DRM will always be cracked (with a sufficiently motivated customer.) AI poisoning will always be cracked (with a sufficiently motivated crawler.)
A huge amount of DRM effort has been spent in the watermarking area, which is similar, but not exactly the same.
This doesn’t stop anyone from viewing or scraping the work, though, so in no way is it DRM. It just causes certain methods of computer interpretation of an image to interpret it in an odd way vs. human viewers. They can still learn from them.
It absolutely is DRM, just a different form than media encryption. It's a purely-digital mechanism of enforcing rights.
It doesn’t enforce any rights. It modifies the actual image. Humans and computers still have equal, open access to it.
15 replies →
Being able to fairly monetise your creative work and put food on the table is a bit rivalrous though, don’t you think?
No, I disagree. There is no principle of the universe or across human civilizations that says that you have a right to eat because you produced a creative work.
The way societies work is that the members of the society contribute and benefit in prescribed ways. Societies with lots of excess production may at times choose to allow creative works to be monetized. Societies without much surplus are extremely unlikely to do so, eg a society with not enough food for everyone to eat in the middle of a famine is extremely unlikely to feed people who only create art; those people will have to contribute in some other way.
I think it is a very modern western idea (less than a century old) that many artists can dedicate themselves solely to producing the art they want to produce. In all other times artists either had day jobs or worked on commission.
> There is no principle of the universe or across human civilizations
Can you list the principles across human civilizations?
3 replies →
> There is no principle of the universe or across human civilizations that says that you have a right to eat because you produced a creative work.
What does that have to do with rivalry? This doesn't dispute the idea that AI is indeed competing with artists. You're just saying artists don't deserve to get paid.
Regardless, some artists will give up but some will simply be more careful with where and how they post their art with tools like these. AI doesn't have a right to the artist's images neither.
1 reply →
No, rivalrous has a specific meaning https://en.wikipedia.org/wiki/Rivalry_(economics)
The tragedy of "your business model is not my problem" as a spreading idea is that while you're right since distribution is where the money is (not creation), intellectual property is de-facto weakened today and IP piracy is widely considered an acceptable thing.
So is sabotaging solutions that would make creative work of the same (or superior) quality more affordable. Your ability to produce expensive illustrations hinders my ability to produce cheap textbooks.
Not everybody equates automated scraping for training models and human experience. Just like any other “data wants to be free” type of discussion, the philosophical and ethical considerations are anything but cut-and-dried, and they’re far more consequential than the technical and economics-in-a-vacuum ones. The general public will quite possibly see things differently than the “oh well, artists— that’s the free market for ya, and you lost” crowd.
You don't copyright ideas, you copyright works. And these artists' productions are works, not abstract ideas, with copyrights, and they are being violated. This is simple law. Why do people have such a hard time with this? Are you the one training the models and you need to find a cognitive escape out of the illegality and wrong-doing of your activities?
It’s not obvious to me that using a copyrighted image to train a model is copyright infringement. It’s certainly not copyright infringement when used to train a human who may end up creating works that are influenced by (but not copies of) the original works.
Now, if the original copyrighted work can be extracted or reproduced from the model, that’s obviously copyright infringement.
OpenAI etc should ensure they don’t do that.
Reproduced to what fidelity? 100%?
If OpenAI's output reproduces a copyrighted image with one pixel changed, is that valid in your view? Where does the line end?
Copyrighted material should never be used for nonacademic language models. "Garbage in, garbage out." All results are tainted.
"But being forced to use non-copyrighted works will only slow things down!"
Maybe that's a good thing, too. Copyright is something every industry has to accept and deal with -- LLMs don't get a "cool tech, do whatever" get-out-of-jail free card.
4 replies →
>Now, if the original copyrighted work can be extracted or reproduced from the model, that’s obviously copyright infringement.
I think there's an important distinction to be made here - "can" be reproduced isn't infringement, only actual reproduction is (and degrees thereof not consisting of sufficiently transformative or fair use).
Trivially a typewriter can reproduce a copyrighted book. Less trivially Google books, with iirc stores the full text of copywrited works has been judged to be legal.
>This is simple law. Why do people have such a hard time with this?
Because this isn’t simple law. It feels like simple infringement, but there’s no actual copying going on. You can’t open up the database and find a given duplicate of a work. Instead you have some abstraction of what it takes to get to a given work.
Also it’s important to point out that nothing in the law is sure. A good lawyer, a sympathetic judge, a bored/interested/contrarian juror, etc can render “settled law” unsettled in an instant. The law is not a set of board game rules.
If the AI were a human and that human made an image that copied substantial elements of another human's creative work after a careful review of the original creator's work, even if it was not an original copy and no archival copy was stored somewhere in the second creator's creative space, I would be concerned about copyright infringement exposure if I were the second (copying) creator.
I'm open to the idea that copyright law might need to change, but it doesn't seem controversial to note that scraping actual creative works to extract elements for an algorithm to generate new works crosses a number of worrying lines.
Have you seen the examples of midjourney reproducing exact frames of Dune, Star Wars etc? With vague prompting not asking for the media property specifically. It's pretty close to querying a database, except if you're asking for something that's not there it's able to render an interpolated result on the fly. Ask it for something that is there however and the model will dutifully pull it up.
3 replies →
> You can’t open up the database and find a given duplicate of a work. Instead you have some abstraction of what it takes to get to a given work.
So distributing a zip file of a copyrighted work subverts the copyright?
> This is simple law.
“One may well ask: ‘How can you advocate breaking some laws and obeying others?’ The answer lies in the fact that there are two types of laws: just and unjust. I would be the first to advocate obeying just laws. One has not only a legal but a moral responsibility to obey just laws. Conversely, one has a moral responsibility to disobey unjust laws. I would agree with St. Augustine that ‘an unjust law is no law at all.’”
Well, fine, but you'll have to claim that copyright is unjust and that you are breaking the law as an act of civil disobedience. The AI corps do not want to take this stance as they also have intellectual property to protect. Classic "have their cake and eat it too" scenario.
Illegality and wrongdoing are completely distinct categories.
I'm not convinced that most copyright infringements are immoral regardless of their legal status.
If you post your images for the world to see, and someone uses that image, you are not harmed.
The idea that the world owes you something after you deliberately shared it with others seems bizarre.
Imagine if every book or advertisement or public conversation you overheard led to future claims that you had unethically learned from public information. It’s such a weird worldview.
(BTW I forbid you from using my comment here in your future reasoning)
> If you post your images for the world to see, and someone uses that image, you are not harmed.
Let me define a few cases of 'uses that image' and see where your line in the sand drops
* If someone used that image as part of an advertising campaign for their product, they are profiting off your work. Are you not harmed?
* If someone used that image and pretended they created it. Are you not harmed?
* If someone used that image and sold it directly. Are you not harmed?
1 reply →
>I'm not convinced that most copyright infringements are immoral regardless of their legal status.
You're right and wrong. You're right because most infringement is from people who can do minimal damage and in fact so more help by giving awareness to your works by sharing. But this is only becsuse copyright it working (most of the time) against corporate entities who don't want to leave any room for legalities to come in.
If copyright ended I'd bet my bottom dollar Disney and all the other billionaires companies would be spamming any and everything that gets moderately popular. And Disney can put advertise the original artist easily.
My statement above is that their activities are illegal and wrong, not that one implies the other. They are illegal because of the copyright violation, and wrong because regardless of what the law says, using the images for training despite the artists' every appeal to the contrary (being vocal about the issue, putting a robots.txt file to avoid scraping, and now using adversarial techniques to protect their work from being stolen) is just moronic. It's like shitting on their front yard when they've asked you a million times not to shit in the front yard, put a sign that says "Please don't shit on my front yard", and sprayed insecticide all over the grass to try to deter you from shitting on the front yard. And yet you still shit on their front yard and even have the balls to argue that there's nothing wrong or illegal about it. This is absolutely insane.
1 reply →
If it were true, then we wouldn't have that great difference in opinions on this topic.
That GP is utterly confused about copyright law is not an opinion.
1 reply →
That may be the law, although we are probably years of legal proceedings away from finding out.
It obviously is not "simple law".
Nothing is being reproduced. Just the ideas being reused.
> ... you cannot “own” an idea.
Let's talk about ownership in a broader sense. In practice, one cannot effectively own (retain possession of) something without some combination of physical capability or coercion (or threat of coercion). Meaning: maintaining ownership of anything (physical or otherwise) often depends on the rule of law.
Then let's use a more precise term that is also present in law: monopoly.
You can't monopolize an idea.
Copyright law is a prescription, not a description. Copyright law demands that everyone play along with the lie that is intellectual monopoly. The effectiveness of that demand depends on how well it can be enforced.
Playing pretend during the age of the printing press may have been easy enough to coordinate, but it's practically impossible here in the digital age.
If we were to increase enforcement to the point of effectiveness, then what society would be left to participate? Surely not a society I am keen to be a part of.
Trying to make sense of the above comment is difficult.
> Copyright law demands that everyone play along with the lie that is intellectual monopoly.
Saying "lie" suggests willful deception. Perhaps you mean "socially constructed"? Combined with "playing pretend" makes it read a bit like a rant.
> Then let's use a more precise term that is also present in law: monopoly.
Ok, in law and economics, the core idea of monopoly has to do with dominant market power that crowds out the existence of others. But your other uses of "monopoly" don't match that. For example, you talk about ideas and "intellectual monopoly". What do you mean?
It seems like some of your uses of "monopoly" are not about markets but instead are closer to the idea of retaining sole ownership.
> If we were to increase enforcement to the point of effectiveness, then what society would be left to participate? Surely not a society I am keen to be a part of.
It appears you've already presupposed how things would play out, but I'm not convinced. What is your metric of effectiveness? A scale is better than some arbitrary threshold.
Have you compared copyright laws and enforcement of the U.S. versus others?
How far would you go: would you say that i.e. society would be better off without copyright law? By what standard?
14 replies →
Kick ass.
I now declare that I own Fortnite.
Where’s my money, Epic?
[flagged]
I don't think we have to compare the process when the end result (another's work being influenced by what they have previously seen) is the same.
Exactly, and the higher the degree of influence the closer to theft. Ai cultists obsessed with stealing property somehow think that’s ok.