A new tool lets artists add invisible changes to the pixels in their art before they upload it online so that if it’s scraped into an AI training set, it can cause the resulting model to break in chaotic and unpredictable ways.
The tool, called Nightshade, is intended as a way to fight back against AI companies that use artists’ work to train their models without the creator’s permission.
[…]
Zhao’s team also developed Glaze, a tool that allows artists to “mask” their own personal style to prevent it from being scraped by AI companies. It works in a similar way to Nightshade: by changing the pixels of images in subtle ways that are invisible to the human eye but manipulate machine-learning models to interpret the image as something different from what it actually shows.
New CAPCHA just dropped.
Lol… I just read the paper, and Dr Zhao actually just wrote a research paper on why it’s actually legally OK to use images to train AI. Hear me out…
He changes the ‘style’ of input images to corrupt the ability of image generators to mimic them, and even shows that the super majority of artists even can’t tell when this happens with his program, Glaze… Style is explicitly not copywriteable in US case law, and so he just provided evidence that the data OpenAI and others use to generate images is transformative which would legally mean that it falls under fair use.
No idea if this would actually get argued in court, but it certainly doesn’t support the idea that these image generators are stealing actual artwork.
So tl;dr he/his team did two things:
- argue the way AI uses content to train is legal
- provide artists a tool to prevent their content being used to train AI without their permission
On the surface it sounds all good, but I can’t help but notice a future conflict of interest for Zhao should Glaze ever become monetized. If it were to be ruled illegal to train AI on content without permission, tools like Glaze would be essentially anti-theft devices, but while it remains legal to train AI this way, tools like Glaze stand to perhaps become necessary for artists to maintain the pre-AI status quo w/r/t how their work can be used and monetized.
There is probably a trivial workaround to this.
It doesn’t even need a work around, it’s not going to affect anything when training a model.
It might make style transfer harder using them as reference images on some models but even that’s fairly doubtful, it’s just noise on an image and everything is already full of all sorts of different types of noise.
The problem is identifying it. If it’s necessary to preprocess every image used for training instead of just feeding it is a model that already makes it much more resources costly
You wouldn’t want to. If you just feed it to the models, then if there are enough of these images to matter the model will learn to ignore the differences. You very specifically don’t want to prevent the model from learning to overcome these things, exactly because if you do you’re stuck with workarounds like that forever, but if you don’t the model will just become more robust to noisy data like this.
There’s trivial workarounds for Glaze, which this is based off of, so I wouldn’t be surprised.
Yes: Train on more images processed by this.
In other words: If the tool becomes popular it will be self-defeating by producing a large corpus of images teaching future models to ignore the noise it introduces.
There are likely easier “quick fixes” while waiting for new models, but this is the general fix that will work against almost any adversarial attack like this.
There might be theoretical attacks that’d be somewhat more difficult to overcome to the extent of requiring tweaks to the models, but given that there demonstrably exists a way of translating text to images that overcomes any such adversarial method that isn’t noticeable to humans, given that humans can, there will inherently always be a way to beat them.
Ooo, this is fascinating. It reminds me of that weird face paint that bugs out facial-recognition in CCTV cameras.
Or the patterned vinyl wraps they used on test cars that interferes with camera autofocus.
this is so dumb and clear it wont work at all. thats not the slightest how ai trains on images.
you would be able to get around this tool by just doing the nft thing and screenshot the image and boom code in the picture is erased.
Here’s the paper: https://arxiv.org/pdf/2302.04222.pdf
I find it very interesting that someone went in this direction to try to find a way to mitigate plagiarism. This is very akin to adversarial attacks in neural networks (you can read more in this short review https://arxiv.org/pdf/2303.06032.pdf)
I saw some comments saying that you could just build an AI that detects poisoned images, but that wouldn’t be feasible with a simple NN classifier or feature-based approaches. This technique changes the artist style itself to something the AI would see differently in the latent space, yet, visually perceived as the same image. So if you’re changing to a different style the AI has learned, it’s fair to assume it will be realistic and coherent. Although maaaaaaaybe you could detect poisoned images with some dark magic tho, get the targeted AI then analyze the latent space to see if the image has been tampered with
On the other hand, I think if you build more robust features and just scale the data this problems might go away with more regularization in the network. Plus, it assumes you have the target of one AI generation tool, there are a dozen of these, and if someone trains with a few more images in a cluster, that’s it, you shifted the features and the poisoned images are invalid
Haven’t read the paper so not sure about the specifics, but if it relies on subtle changes, would rounding color values or down sampling the image blur that noise away?
Wondering the same thing. Slight loss of detail but still successfully gets the gist of the original data.
For that matter, how does the poisoning hold up against regular old jpg compression?
Eta: read the paper, they account for this in section 7. It seems pretty robust on paper, by the time you’ve smoothed out the perturbed pixels, youve also smoothed out the image to where the end result is a bit of a murky mess.
Trying to detect poisoned images is the wrong approach. Include them in the training set and the training process itself will eventually correct for it.
I think if you build more robust features
Diffusion approaches etc. do not involve any conscious “building” of features in the first place. The features are trained by training the net to match images with text features correctly, and then “just” repeatedly predict how to denoise an image to get closer to a match with the text features. If the input includes poisoned images, so what? It’s no different than e.g. compression artifacts, or noise.
These tools all try to counter models trained without images using them in the training set with at most fine-tuning, but all they show is that models trained without having seen many images using that particular tool will struggle.
But in reality, the massive problem with this is that we’d expect any such tool that becomes widespread to be self-defeating, in that they become a source for images that will work their way into the models at a sufficient volume that the model will learn them. In doing so they will make the models more robust against noise and artifacts, and so make the job harder for the next generation of these tools.
In other words, these tools basically act like a manual adversarial training source, and in the long run the main benefit coming out of them will be that they’ll prod and probe at failure modes of the models and help remove them.
Just to start with, not very experienced with neural networks at all beyond messing with openCV for my graduation project.
Anyway, that these countermeasures expose “failure modes” in the training isn’t a great reason to stop doing this, e.g. scammers come up with a new technique, we collectively respond with our own countermeasures.
If the network feedbacks itself, then cool! It has developed its own style, which is fine. The goal is to stop people from outright copying existing artists style.
It doesn’t need to “develop its own style”. That’s the point. The more examples of these adversarial images are in the training set, the better it will learn to disregard the adversarial modifications, and still learn the same style. As much as you might want to stop it from learning a given style, as long as the style can be seen, it can be copied - both by humans and AI’s.
There’s a lot of interesting detail to your side of the discussion I may not yet have the knowledge of. How does the eye see? We find edges, gradients, repeating patterns which become textures, etc etc… But our systems can be misdirected, see the blue/yellow dress for example. NNsbhave the luxury of being rapidly iterated I guess, compared to our lifespans.
I’m asking questions I don’t know answers to here: if the only source of input data for a network is subtly corrupted, won’t that guarantee corrupted output as well? I don’t see how one can “train out” the corruption which misdirects the network without access to some pristine data.
Don’t get me wrong, I’m not naive enough to believe this is foolproof, but I do want to understand why this technique doesn’t actually work, and by extension better understand how training a nn actually works.
if the only source of input data for a network is subtly corrupted, won’t that guarantee corrupted output as well?
We have to distinguish between different kinds of “corruption”, here. What you seem to be describing is “if we only feed the model data from rule34, will it ever learn proper human anatomy” and the answer is no, it won’t. You’ll have to add data which narrows the range of body proportions from cartoonish to, well, real. That’s an external source of corruption: Feeding it bad data (for your own definition of “bad”). Garbage in, garbage out.
The corruption that these adversarial models are exploiting though is inherent in the model they’re attacking. Take… ropes and snakes and cats (or, generally, mammals). Good example: It is incredibly easy for a cat to mistake a rope for a snake – it looks exactly the same to the first layers of the visual cortex and evolution would rather have the cat jump away as soon as possible than be bitten, and it doesn’t hurt to jump away from a rope (even though the cat might end up being annoyed or ashamed (yes cats can 110% be self-conscious different story)), so when there’s an unexpected wiggly shape the first layers directly tell the motor cortex to move, short-circuiting any higher processing.
That trait has been written into the network by evolution, very similar to how we train AI models – conceptually, that is: In both cases the network gets trained for fitness for a purpose (the implementation details are indeed rather different but also irrelevant):
What those adversarial models do kinda looks like this: Take a picture of a rope. Now randomly shift pixels to make the rope subtly more snake-like until you get your cat to jump as reliably as possible, in as many different situations as possible, e.g. even if they’re expecting it and staring straight at it. Sell the product for a lot of money. People start posting pictures of ropes, rope manufacturers adjust their weaving patterns. Other cats see those pictures and ropes, some jump, and others only feel a bit, or a lot, uneasy. The ones that jump will not be able to procreate, any more, being busy jumping, while the uneasy ones will continue to evolve. After a couple of generations no cat cares about those ropes with shifted pixels any more.
Whether that trains general immunity against adversarial attacks – I wouldn’t be so sure. It very likely will make the rope/snake distinction more accurate. But even if it doesn’t build general immunity, it’s an eternal cat and mouse game and no artist will be willing to continue paying for that kind of software when it’s going to get defeated within days, anyway, because that’s just how fast we can evolve models.
Oh. Back to the definition of corruption: If all the pictures of rope that our models ever see have shifted pixels then it’s just going to assume that is the norm, and distinguish it from snakes because the tags say “rope” in one case, and “snake” in the other. The original un-shifted pictures probably won’t be an adversarial attack because they’re not a product of trying to get cats to jump.
Quick iteration is definitely the big thing. (The eye is fun because it’s so “badly designed” - we’re stuck in a local maxima that just happens to be “good enough” for us to not overcome the big glaring problems)
And yes, if all the inputs are corrupted, the output will likely be too. But 1) they won’t all be, and as long as there’s a good mix that will “teach” the network over time that the difference between a “corrupted cat” and an “uncorrupted cat” are irrelevant, because both will have most of the same labels associated with them. 2) these tools work by introducing corruption that humans aren’t meant to notice, so if the output has the same kind of corruption it doesn’t matter. It only matters to the extent the network “miscorrupts” the output in ways we do notice enough so that it becomes a cost drag on training to train it out.
But you can improve on that pretty much with feedback: Train a small network to recognize corruption, and then feed corrupted images back in as negative examples to teach it that those specific things are particularly bad.
Picking up and labelling small sample sets of types of corruption humans will notice is pretty much the worst case realistic effect these tools will end up having. But each such countermeasure will contribute to training sets that make further corruption progressively harder. Ultimately these tools are strictly limited because they can’t introduce anything that makes the images uglier to humans, and so you “just” need to teach the models more about the limits of human vision, and in the long run that will benefit the models in any case.
deleted by creator
“I can tell this is toxic by the pixels.”
“We like to call them poison pixels.”
What a dumb solution to a problem that doesn’t need a solution. The problem isn’t AI, it’s the lack of understanding for the tech that has people thinking AI is theft.
Is it not theft? These “AI” are trained on other people’s work, often without their knowledge or permission.
This is why I think people don’t know what they are talking about.
You can look at a picture from an artist without it being considered theft, so are your memories and impressions theft? That’s what training data does, it teaches AI what something looks like, with many samples. It’s literally what your brain does, the way you see multiple dogs and know what a dog looks like is the same way that AI trains pattern recognition.
It’s completely reasonable and desirable to have AI consume all available images, regardless of copyright the way your eyes and brain can do the same. Training data isn’t theft no more than going to a museum and looking at art is theft.
This take that this is bad is completely unhinged and indicates people don’t understand AI.
I’d be careful with claiming who does and does not understand things.
First of all, a person can’t go to a museum, see a piece of art then go home and reproduce that art or style. Given enough time, sure they might be able to learn to replicate the style. Those that are particularly good at reproduction might even become forgers which is a crime.
Second, these llms aren’t AI. They can’t think in terms of how a living being can, only regurgitate information. They’re glorified search engines in a way.
Lastly, I can assume that you aren’t a creative person. You probably type in some prompt to an image generator and think “I made this”. It’s easier for someone like you to overlook issues because they don’t effect you because you lack depth, which I know is hard to accept. Maybe one day you’ll gain some insight into your own lack of understanding… But I doubt it.
I used to be a musician, I also used to paint. I think my thought processes are no more complex than most computers, and I genuinely don’t believe human creativity is special even a little bit, like consciousness, it’s a subjective illusion.
I do not believe in things like copyright, or intellectual property, or even ownership of these things, I think these things should be collectively owned by society.
I don’t disagree with you from lack of experience, I disagree from fundamentally different ideological underpinnings.
I believe there is nothing special about human perception and experience, and I can see the ways that technology maps near perfectly to the way we think. AI shouldn’t be limited, it should replace us.
Okie dokie, doc. If you think the human brain isn’t “special” then I don’t know what to tell you.
Also, you can’t know how we think when we as a species don’t know, but you being the smartest person in the room is clearly very important to you so I’ll leave you to it!
I remember in the early 2010s reading an article like this one on openai.com talking about the dangers of using AI for image search engines to moderate against unwanted content. At the time the concern was CSAM salted to prevent its detection (along with other content salted with CSAM to generate false positives).
My guess is since we’re still training AI with pools of data-entry people who tag pictures with what they appear to be, so that AI reads more into images than their human trainers (the proverbial man inside the Iron Turk).
This is going to be an interesting technology war.
Obviously this is using some bug and/or weakness in the existing training process, so couldn’t they just patch the mechanism being exploited?
Or at the very least you could take a bunch of images, purposely poison them, and now you have a set of poisoned images and their non-poisoned counterparts allowing you to train another model to undo it.
Sure you’ve set up a speedbump but this is hardly a solution.
Obviously this is using some bug and/or weakness in the existing training process, so couldn’t they just patch the mechanism being exploited?
I’d assume the issue is that if someone tried to patch it out, it could legally be shown they were disregarding people’s copyright.
It isn’t against copyright to train models on published art.
Explain
In order to violate copyright you need to copy the copyrighted material. Training an AI model doesn’t do that.
The general argument legally is that the AI has no exact memory of the copyrighted material.
But if that’s the case, then these pixels shouldn’t need be patched. Because it wouldn’t remember the material that spawned them.
Is just the argument I assume would be used.
What is “patching pixels” and who would do it?
Is that not answered in the original article?
It’s like training an artist who’s never seen a banana or a fire hydrant, by passing them pictures of fire hydrants labelled “this is a banana”. When you ask for a banana, you’ll get a fire hydrant. Correcting that mistake doesn’t mean “undoing pixels”, it means teaching the AI what bananas and fire hydrants are.
Well, I guess we’ll see how that argument plays in court. I don’t see how it follows, myself.
Obviously, with so many different AIs, this can not be a factor (a bug).
If you have no problem looking at the image, then AI would not either. After all both you and AI are neural networks.
An AI don’t see the images like we do, an AI see a matrix of RGB values and the relationship they have with each other and create an statistical model of the color value of each pixel for a determined prompt.
deleted by creator
The neural network of a human and of an AI operate in fundamentally different ways. They also interact with an image in fundamentally different ways.
I would not call it “fundamentally” different at all. Compared to, say, regular computer running non-neural network based program, they are quite similar, and have similar properties. They can make a mistake, hallucinate, etc.
As a person who has done machine learning, and some ai training and who has a psychotic disorder I hate they call it hallucinations. It’s not hallucinations. Human hallucinations and ai hallucinations are different things. One is based of limited data , bias, or a bad data set with builds a fundamentally bad neural network connection which can be repaired. The other is something that can not be repaired, you are not working with bad data, your brain can’t filter out data correctly and you are building wrong connections. It’s like an overdrive of input and connections that are all wrong. So you’re seeing things, hearing things, or believing things that aren’t real. You make logical leaps that are irrational and not true and reality splits for you. While similarities exist, one is because people input data wrong, or because they cleaned it wrong, or didn’t have enough. And the other is because the human brain has wiring problem caused by a variety of factors. It’s insulting and it also humanizes computers to much and degrades people with this illness.
As I understand, healthy people hallucinate all the time, but in different sense, non-psychiatric sense. It is just healthy brain has this extra filter that rejects all hallucinations that do not correspond to the signal coming from reality, that is our brain performs extra checks constantly. But we often get fooled if we do not have checks done correctly. For example, you can think that you saw some animal, while it was just a shade. There is even statement that our perception of the world is “controlled hallucination” because we mostly imagine the world and then best fit it to minimize the error from external stimuli.
Of course, current ANNs do not have such extensive error checking, thus they are more prone to those “hallucinations”. But fundamentally those are very similar to what we have in those “generative suggestions” our brain generates.
Those aren’t quite the same as a hallucination. We don’t actually call them hallucinations. Hallucinations are a medical term. Those are visual disturbances not “controlled hallucinations”. Your brain filtering it out and the ability to ignore it makes it not a hallucination. It’s hallucinations in a colloquial sense not medical.
Fundamentally AI is not working the same, you are having a moment of where a process from when in the past every shadow was a potential danger so seeing a threat in the shadow first and triggering fight or flight is best for you as a species. AI has no fight or flight. AI has no motivation, AI just had limited, bad, or biased data that we put there and spits out garbage. It is a computer with no sentience. You are not really error checking, you are processing more information, or reassessing once the fight/flight goes down. AI doesn’t have more information to process.
Many don’t see people with psychotic disorders as equal people. They see them as dangerous, and and people to be locked away. They use their illnesses and problems as jokes and slurs. Using terms for their illness in things like this only adds to their stigma.
You are arguing about terminology use. Please google “controlled hallucinations” to see how people use the term in non-psychiatric way.
No! It’s not using an internal exploit, it’s rather about finding a way to visually represent almost the same image, but instead using latent features with different artists (e.g, which would confuse a dreambooth+lora training), however, the method they proposed is flawed, I commented more on https://lemmy.world/comment/4770884
I am waiting for the day that some obsessed person starts finding ways to do like code injection in pictures.
Invisible changes to pixels sound like pure BS to me. I’m sure others know more about it than i do but I thought pixels were very simple things.
“Invisible changes to pixels” means “a human can’t tell the difference with a casual glance” - you can still embed a shit-ton of data in an image that doesn’t look visually like it’s been changed without careful inspection of the original and the new image.
If this data is added in certain patterns it will cause ML models trained against the image to draw incorrect conclusions. It’s a technical hurdle that will slow a casual adversary, someone will post a model trained to remove this sometime soon and then we’ll have a good old software arms race and waste a shit ton of greenhouse emissions adding and removing noise and training ever more advanced models to add and remove it.
You can already intentionally poison images so that image recognition draws incorrect conclusions fairly easily, this is the same idea but designed to cripple ML model training.
Pixels are very simple things, literally 3-5 3 digit numbers.
But pixels mean little too a generative AI - it’s all about relationship between pixels. All AI are high dimensional shapes right now… If you break up the shape strategically, it’ll poison the image
Will this poison pill work? Probably, for at least a while…
I’m sure others know more about it than i do but I thought pixels were very simple things.
You’re right, in that pixels are very simple things. However, you and I can’t tell one pixel from another in an image, and at the scale of modern digital art (my girlfriend does hers at 300dpi), shifting a handful of pixels isn’t going to make much of a visible difference to a person, but a LLM will notice them.
An AI model will “notice them” but ignore them if trained on enough copies with them to learn that they’re not significant.
LLM is the wrong term. That’s Large Language Model. These are generative image models / text-to-image models.
Truthfully though, while it will be there when the image is trained, it won’t ‘notice’ it unless you distort it significantly (enough for humans to notice as well). Otherwise it won’t make much of a difference because these models are often trained on a compressed and downsized version of the image (in what’s called latent space)
A pixel has a binary representation. All of the significant bits for the pixel may not not be needed to display the color of that pixel so there is often excess that can be used or modified. A person wouldn’t see it but an AI reading just the binary would.
deleted by creator
How is training AI with art on the web different to a person studying art styles? I’d say if the AI is being monetized in some capacity, then sure maybe there should be laws in place. I’m just hard-pressed to believe that anyone can have sole control of anything once it gets on the Internet.
How is training AI with art on the web different to a person studying art styles?
Human brains clearly work differently than AI, how is this even a question?
The term “learning” in machine learning is mainly a metaphor.
Also, laws are written with a practical purpose in mind - they are not some universal, purely philosophical construct and never have been.
Human brains clearly work differently than AI, how is this even a question?
It’s not all that clear that those differences are qualitatively meaningful, but that is irrelevant to the question they asked, so this is entirely a strawman.
Why does the way AI vs. the brain learn make training AI with art make it different to a person studying art styles? Both learn to generalise features that allows them to reproduce them. Both can do so without copying specific source material.
The term “learning” in machine learning is mainly a metaphor.
How do the way they learn differ from how humans learn? They generalise. They form “world models” of how information relates. They extrapolate.
Also, laws are written with a practical purpose in mind - they are not some universal, purely philosophical construct and never have been.
This is the only uncontroversial part of your answer. The main reason why courts will treat human and AI actions different is simply that they are not human. It will for the foreseeable future have little to do whether the processes are similar enough to how humans do it.
Now you’re just cherry picking some surface-level similarities.
You can see the difference in the process in the results, for example in how some generated pictures will contain something like a signature in the corner, simply because it resembles the training data - even though there is no meaning to it. Or how it is at least possible to get the model to output something extremely close to the training data - https://gizmodo.com/ai-art-generators-ai-copyright-stable-diffusion-1850060656.
That at least proves that the process is quite different to the process of human learning.
The question is how much those differences matter, and which similarities you want to focus on.
Human learning is similar in some ways, but greatly differs in other ways.
The fact that you’re picking and choosing which similarities matter and which don’t is just your arbitrary choice.
You can see the difference in the process in the results, for example in how some generated pictures will contain something like a signature in the corner
If you were to train human children on an endless series of pictures with signatures in the corner, do you seriously think they’d not emulate signatures in the corner?
If you think that, you haven’t seen many children’s drawings, because children also often pick up that it’s normal to put something in the corner, despite the fact that to children pictures with signatures is a tiny proportion of visual input.
Or how it is at least possible to get the model to output something extremely close to the training data
People also mimic. We often explicitly learn to mimic - e.g. I have my sons art folder right here, full of examples of him being explicitly taught to make direct copies as a means to learn technique.
We just don’t have very good memory. This is an argument for a difference in ability to retain and reproduce inputs, not an argument for a difference in methods.
And again, this is a strawman. It doesn’t even begin to try to answer the questions I asked, or the one raised by the person you first responded to.
That at least proves that the process is quite different to the process of human learning.
Neither of those really suggests that all (that diffusion is different to humans learn to generalize images is likely true, what you’ve described does not provide even the start of any evidence of that), but again that is a strawman.
There was no claim they work the same. The question raised was how the way they’re trained is different from how a human learns styles.
I appreciate your responses, thank you!
I work in AI and I believe it is different. Society is built to distribute wealth, so that everyone can live a decent life. People and AI should be treated differently in front of the law. Also, non-commercial, open source AI should be treated differently than commercial or closed source models
Society is built to distribute wealth, so that everyone can live a decent life.
As a goal, I admire it, but if you intend this as a description of how things are it’d be boundlessly naive.
That’s absolutely not how it is now, just the goal we should set for ourselves. A goal I believe we should consider when regulating AI
To me, that’s not an argument for regulating AI, though, because most regulation we can come up with will benefit those with deep enough pockets to buy themselves out of the problem, while solving nothing.
E.g. as I’ve pointed out in other debates like this, Getty Images has a market cap of <$2bn. OpenAI may have had a valuation in the $90bn range. Google, MS, Adobe all also have shares prices that would trivially allow them to purchase someone like Getty to get ownership of a large training set of photos. Adobe already has rights to a huge selection via their own stock service.
Bertelsmann owns Penguin Random-House and a range ofter publishing subsidiaries. It’s market cap is around 15 billion Euro. Also well within price for a large AI contender to buy to be able to insert clauses about AI rights. (You think authors will refuse to accept that? All but the top sellers will generally be unable to afford to turn down a publishing deal, especially if it’s sugar-coated enough, but they also sit on a shit-ton of works where the source text is out-of-copyright but they own the right to the translations outright as works-for-hire)
That’s before considering simply hiring a bunch of writers and artists to produce data for hire.
So any regulation you put in place to limit the use of copyrighted works only creates a “tax” effectively.
E.g. OpenAI might not be able to copy artist X’s images, but they’ll be able to hire artist Y on the cheap to churn out art in artist X’s style for hire, and then train on that. They might not be able to use author Z’s work, but they can hire a bunch of hungry writers (published books sells ca 200 copies on average; the average full time author in the UK earns below minimum wage from their writing) as a content farm.
The net result for most creators will be the same.
Even wonder why Sam Altmann of OpenAI has been lobbying about the dangers of AI? This is why. And its just the start. As soon as these companies have enough capital to buy themselves access for data, regulations preventing training on copyrighted data will be them pulling up the drawbridge and making it cost-prohibitive for people to build open, publicly accessible models in ways that can be legally used.
And in doing so they’ll effectively get to charge an “AI tax” on everyone else.
If we’re going to protect artists, we’d be far better off finding other ways of compensating them for the effects, not least because it will actually provide them some protection.
UBI is the known solution to protect workers. Solution is there, people aren’t ready for it
As long as people aren’t ready for it, then it doesn’t solve the immediate problem that needs to be solved today.
Lol.
How does UBI break trademark and copyright law (and therefore legal cases)?
Do you really think the current power brokers will suddenly sit in their hands and stop trying to (mostly successfully) control as much as they can?
I agree that the training isn’t fundamentally different, but that monetization of the output has to be controlled. The big difference between AI and humans is the speed with which they create - you have to employ an army of humans to match the output of a couple of GPUs. For noncommercial projects this is amazing. For commercial projects, it destroys the artists livelihoods.
But this simply means that training shouldn’t be controlled, inference in commercial contexts should be.
The real issue comes in ownership of the AI models and the vast amount of labor involved in the training data. It’s taking what is probably hundreds of thousands of hours of labor in the form of art and converting it into a proprietary machine, all without compensating the artists involved. Whether you can make a comparison to a human studying art is irrelevant, because a corporation can’t own an artist, but they can own an AI and not have to pay it.
Until the law catches up with the technology, people need ways of protecting themselves.
I agree, and I wonder if the law might be kicked into catching up quicker as more companies try to adopt these tools and inadvertently infringe on other companies’ copyrighted material. 😅
Disagree. It’s only unethical if you use it to generate the artist’s existing pieces and claim it as yours.
deleted by creator
I don’t see how AI training couldn’t be considered transformative as the whole idea is to consume input, break it down into data, and output something new. The way I’m understanding what you’re saying is like this: Instead of only paying royalties when I try to monetize a cover song, I’d have to pay every time I practiced it.
deleted by creator
I don’t understand how you’re separating the the generated artworks from the AI that’s generating the work, but I do see your point. If a company puts out a tool for free I don’t think they should be on the hook for someone using that and creating a product. At the end of it all though, I think whoever has made any hard financial gains should should payout whoever contributed.
I don’t see a problem with it training on all materials, fuck copyright. I see the problem in it infringing on everyone’s copyright and then being proprietary, monetized bullshit.
If it trains on an open dataset, it must be completely and fully open. Everything else is peak capitalism.
You’re not owed nor entitled to an artist’s time and work for free.
I am perfectly entitled to type random stuff into google images, pick out images for a mood board and some as reference, regardless of their copyright status, thank you. Studying is not infringement.
It’s what every artist does, it’s perfectly legal, and what those models do is actually even less infringing because they’re not directly looking at your picture of a giraffe and my picture of a zebra when drawing a zebra-striped giraffe, they’re doing it from memory.
Art takes effort. You’re not entitled to that for free.
And if you think that working with AI does not take effort you either did not try, or don’t have an artistic bone in your body. Randos typing “Woman with huge bazingas” into an UI and hitting generate don’t get copyright on the output, rightly so: Not just did they not do anything artistic, they also overlook all the issues with whatever gets generated because they lack the trained eye of an artist.
Of course not, it’s the artists decision to put it on the internet for free.
Technically that’s the root of the issue. This does not grant a license to everyone who looks at it, but if a license is required to train a model is unclear and currently discussed in court.
The problem is the only way for artists to get people to see and eventually buy their art or commissions is to post some of their work publicly. Historically you would go out on the street and set up a stall, now social media is our digital street. Galleries don’t take everyone, having the ability to even get a meeting with one is difficult without the right connections. Most artists are never successful enough to completely live off their art, if they can make any money at all it is great for them. Then along comes an AI model that takes their work because it’s on the internet scrapes it into its training set and now any chance they had in an over saturated market is even smaller, because hey, I can just do this with AI. This idea that copyright and IP shouldn’t exist at all is kinda absurd. Would you just go through a street art walk, take high res photos of every picture they have on display, not take any business cards, and when they ask what you are doing, go “it’s ok, I’m training an AI data model so people can just make work that looks exactly like this. They shouldn’t have to ever buy from you. Capitalism is a joke. Bye!” The art walk was free, but it was also a sales pitch, because that’s how the art world works. You are hoping to get seen, that someone likes it enough to buy, and maybe buy more.
This idea that copyright and IP shouldn’t exist at all is kinda absurd
I don’t hold this opinion at all.
I’m just saying that there are uses for which you don’t need a license. Say, visiting an art exhibition and then going home and trying to draw similar pictures. Wether AI training falls into this category or instead requires a license is currently unclear.
Btw, two spaces before the line break
Creates the spacing you want.As an artist who studies data science, I would say doing art and generating art are an entirely different process. AI has no reference outside of the information we give it. It had no real understanding of lighting, spacial awareness. We can tell it every tank is a cat, every flashlight is a pig and it will never question it. If we tell a toddler that every tank is a cat, they may call a tank a cat, but they will never think a that “cat” is a house pet. They will never think that “pig” will oink or be turned into steaks. An AI however would if your language conventions were the same in the prompt.
If you go to the art walk and go home and try to recreate a style, you were inspired. If an AI model is trained on many styles and you tell it “portrait, woman, Van Gogh style, painterly, blue tones” then do you understand what you asked for? Was the ai inspired by Van Gogh? Did the ai study his techniques? No. It broke down his art pixel by pixel, rearranged it in a filter styled overlay over a woman, most likely a young woman-because of algorithmic bias which has been studied- in shades of blue. Humans take the time to study the why, the how. Ai does not. Humans are not just meat robots.
I should say I’m not against AI art. I’m against gathering against consent. If it was opt in, or if there was some type of pay for program that would be fine. Even if it was pennies each month. But the fact that they scrape without consent. Or are now going back and adding it into TOS where it never was before feels scummy. AI art has a place, and is a helpful tool. But it’s not a replacement for artists, it has many flaws still, that might never be worked out.
Thank you for helping me with line break.
This idea that copyright and IP shouldn’t exist at all is kinda absurd.
For the majority of human existence, that was the default.
Copyright exists as an explicit tradeoff between the rights of the public to be able to do as they please with stuff introduced into the public sphere, and a legal limitation infringing on the publics liberty for a limited time for the purpose of encouraging the creation of more works for the public benefit. It was not introduced as some sort of inherent right, but as a trade between the public and creators to incentivise them.
Stripping it away from existing artists who has come to depend on it without some alternative would be grossly unfair, but there’s nothing absurd about wanting to change the bargain over time. After all, that has been done many times, and the copyright we have now is vastly different and far more expansive and lengthy than early copyright protection.
Personally, I’d be in favour of finding alternative means of supporting creators and stripping back copyright as a tradeoff. The vast majority of creators earn next to nothing from their works; only a very tiny minority makes a livable wage of art of any form at all, and of the rest the vast majority of profits take place in a very short period of initial exploitation of a work, so we could allow the vast majority to earn more from their art relatively cheaply, and affect the rest to a relatively limited degree, while benefiting from the reduced restrictions.
I agree that copyright lasts far too long, but the idea I can post a picture today, and in a hour it’s in an AI model without my consent bothers me. Historically there was a person to person exchange. But now we are so detached from it all I don’t think we can have that same affordance of no types of protections. I’m not saying one person can solve this. But I don’t see UBI or anything like that ever happening. As a person who has lived on disability most of their life, people don’t like to share their wealth with anyone for any reason. I’ve never been able to sell art for a living and am now going to school for data science. So I know about both ends of this. Just scraping without consent is unethical and many who do this have no idea about the art world or how artist create in general.
I doesn’t need to be full on UBI. In a lot of countries grants mechanisms and public purchasing mechanisms for art already make up a significant proportion of income for artists. Especially in smaller countries, this is very common (more so for literary works, movies and music where language provides a significant barrier to accessing a bigger audience, but for other art too). Imagine perhaps a tax/compulsory licensing mechanism that doesn’t stop AI training but instead massively expands those funding sources for people whose data are included in training sets.
This is not stoppable, not least because it’s “too cheap” to buy content outright.
I pointed out elsewhere that e.g. OpenAI could buy all of Getty Images for ~2% of their currently estimated market cap based on a rumoured recent cash infusion. Financing vast amounts of works for hire just creates a moat for smaller players while the big players will still be able to keep improving their models.
As such it will do nothing to protect established artists, so we need expansion of ways to fund artists whether or not inclusion of copyrighted works in training sets becomes restricted.
This is already a concept in the AI world and is often used while a model is being trained specifically to make it better. I believe it’s called adversarial training or something like that.
Its called adversarial attack, this is an old video (5 years) explaining how it works and how you can potentially do it charging just one pixel on the image.
Here is an alternative Piped link(s):
https://piped.video/SA4YEAWVpbk?si=xObPveXTT2ip5ICG
Piped is a privacy-respecting open-source alternative frontend to YouTube.
I’m open-source; check me out at GitHub.
No, that’s something else entirely. Adversarial training is where you put an ai against a detector AI as a kind of competition for results.
