Comment on This new data poisoning tool lets artists fight back against generative AI
kayrae_42@lemmy.world 1 year agoI don’t see these grants or public funding ever covering a private company for one. And for two, I don’t see AI art ever actually getting to the point where it fully replaces artists. As of right now it is good. But it doesn’t understand space or lighting at all. Because of how AI works I’m not sure it ever will. Because it is trained to make a homogeneous rendering of what you are looking for, even if you use a base image, most people have an image that is lit heavily in the front, but because of this it never is able to render shadows correctly. Unless they hire people who are artist or art critics to finely train the data set, which I doubt they will, then the more you look the more uncanny valley the images get. They also have a hard bias in all of their images they generate. Which is difficult to overcome.
AI is an amazing tool, but it is a poor replacement in total. The people who act like it is a total replacement are like the people who in 2015 told us self driving cars were just one year away, and have been saying it every year since. Maybe when quantum computing becomes the standard for every person AI will be able to. But there is just a fundamental misunderstanding of art, artistic process, how art get made people seem to have.
Open AI might be sitting on Microsoft money, but how many other companies has Microsoft gobbled up over the years? Open AI if it starts to struggle will just fall under the Microsoft umbrella and become part of its massive conglomerate, integrated into it. Where are our AR goggles that we are supposed to all be wearing, Microsoft and Google both had those? So many projects grow and die with multiple millions thrown at them. All end up with crazy valuations based on future consumer usage. As we all can’t even afford rent.
There is also this idea that people wouldn’t willing contribute if just asked. The problem is no one has even asked. Hugging Face is an open source distro people willingly contribute to. And so many people upload images to Creative Commons which could be used. I’ve done it with many of my photos which I have no problem being used in a data set, for commercial use even. But my commercial images, no please. The idea that you can’t train smaller models on the vast array of Creative Commons images and public domain, you absolutely can. You can also ask people to contribute to your data set and give credit to them. A lot of people are angry at lack of credit.
There is no reason for any of this to be private enterprise if they are going to blatantly steal copyright images when sources like Creative Commons exists, not give any credit to the people they steal from, and sometime even steal from places they shouldn’t even have access to.
vidarh@lemmy.stad.social 1 year ago
Companies are by far the largest recipients of public funding for art in many countries and sectors. Especially for e.g. movie production in smaller languages, but also in other sectors.
I do agree it won’t fully replace artists, but not because it won’t get to the point where it can be better than everyone, but because a huge part of art is provenance. A “better Mona Lisa” isn’t worth anything, while the original is priceless, not because a “better” one isn’t possible, but because it’s not painted by Da Vinci.
But that will only help an even narrower sliver than the artists who are making good money today.
It will take time, but AI will eat far more fields than art, and we haven’t even started to see the fallout yet.
Diffusion models are not trained “for” anything other than matching vectors to denoising to within your own tolerance levels of matching to what you are looking for. Accordingly, you’ll see a whole swathe of models tuned on more specific types of imagery, and tooling to more precisely control what they generate. The “basic” web interfaces are just scratching the surface of what you can do with e.g. Controlnet and the like. It will take time before they get good enough, sure. They are also only 2 years old, and people have only been working on tooling around then for much less than that.
OpenAI is just one of many in this space already. They are in the lead for LLMs, that is text-based models. But even that lead is rapidly eroding. They don’t have any obvious lead for diffusion models for images. Having used several, it was first with the recent release of DallE 3 that it got “good enough” to be competitive.
At the same time there are now open models getting close enough to be useful, so even if every AI startup in the world collapsed this won’t go away.
That’s fine, but that doesn’t fix the financial challenge.
kayrae_42@lemmy.world 1 year ago
So what you are saying is open ai should get the public grants for artists to give to artists?
I understand it isn’t trained for anything, I have done training with them. The training leads to homogeneous outcomes. It had been studied as well. You can look it up.
Dall-e 3 still isn’t good enough to be competitive. It is too uncanny valley. I’m not saying people have to be the masters. I don’t know where you get that from, every one who touts this tech always goes to that. It is a tool that can be useful, but it is not a replacement.
Asking and crediting would go a long way to help fix the financial challenge. Because it is a start to adding a financial component. If you have to credit someone there becomes an obligation to that person.
vidarh@lemmy.stad.social 1 year ago
No. What in the world gave you that idea? I’m saying artists or companies employing artists should get grants, just like is the case for a large number of grants now. I’m saying I’d like to see more of that to compensate for the effects being liberal about copyright would have.
There is no “the training”. There are a huge range of models trained with different intent producing a wide variety in output to the point that some produces output that others will just plain refuse.
Dall-E 3 isn’t anywhere near leading edge of diffusion models. It’s OpenAI playing catch up. Now, neither Midjourney or Firefly, nor any of the plethora of Stable Diffusion derived models are good enough to be competitive with everyone without significant effort either, today, but that is also entirely irrelevant. Diffusion models are two years old, and the pace of the progress have been staggering, to the point where we e.g. already have had plenty of book-covers and the like using them. Part of the reason for that is that you can continue training of a decent diffusion model even on a a somewhat beefy home machine and get a model that fits your needs better to an extent you can’t yet do with LLMs.
If there is a chance crediting someone will lead to a financial obligation, people will very quickly do the math on how cheaply they can buy works for hire instead. And the vast bulk of this is a one-off cost. You don’t need to continue adding images to teach the models already known thing, so the potential payout on the basis of creating some sort of obligation. Any plan for fixing the financial challenge that hinges on copyright is a lost cause from the start because unless it’s a pittance it creates an inherent incentive for AI companies to buy themselves out of that obligation instead. It won’t be expensive.
kayrae_42@lemmy.world 1 year ago
I feel like you are one of the people who feel that AI is just going to be the future with no real problems to anyone who matters. We can’t stop it, we can’t regulate it in any way whatever; and people should just move out of the way, give up and if they can’t find a place in the new world, die already. Artists don’t matter, writers don’t matter and anyone impacted by this new system doesn’t matter. The algorithm is all that matters.
Because I don’t use the exact correct wording, I use a short hand that is easier for my brain to remember, and you are pedantic, I can’t know anything about LLMs, machine learning or anything about this. Because I don’t say it has a training set of a large model of images that are tagged in specific ways that they can take out antagonistic images or images that create artifacts and refine the model in appropriate ways. You therefore throw out the idea that bias exists due to tagging systems.
Honestly I don’t care if you don’t think I know anything about this. You are a stranger on the internet and this conversation has gone on too long.