Artificial Intelligence
- Started
- Last post
- 1,341 Responses
- jonny_quest_lives0
Exhibit A in the upcoming IP lawsuits Midjourney will likely face... not from people like you or I because we are largely irrelevant but IP giants like Disney aren't going to let a commercial product crap out Mickey renders without their cut.
Forbes interview with David Holz and anyone who has done any sort of commercial work at any time in there lives can see the red flags in all of his statements in regards to Copyright law and IP rights,
"Does Midjourney’s license allow for commercial use of imagery generated by the platform?
Yes. But if you’re working for a company bigger than a million dollars in annual revenue, we ask that you buy a corporate license.
How was the dataset built?
It’s just a big scrape of the Internet. We use the open data sets that are published and train across those. And I’d say that’s something that 100% of people do. We weren’t picky. The science is really evolving quickly in terms of how much data you really need, versus the quality of the model. It’s going to take a few years to really figure things out, and by that time, you may have models that you train with almost nothing. No one really knows what they can do.
Did you seek consent from living artists or work still under copyright?
No. There isn’t really a way to get a hundred million images and know where they’re coming from. It would be cool if images had metadata embedded in them about the copyright owner or something. But that's not a thing; there's not a registry. There’s no way to find a picture on the Internet, and then automatically trace it to an owner and then have any way of doing anything to authenticate it.
Can artists opt out of being including in your data training model?
We’re looking at that. The challenge now is finding out what the rules are, and how to figure out if a person is really the artist of a particular work or just putting their name on it. We haven’t encountered anyone who wants their name taken out of the data set that we could actually find in the data set.
Can artists opt out of being named in prompts?
Not right now. We’re looking at that. Again, we’d have to find a way to authenticate those requests, which can get complicated."
- Opting out is possible in some places now https://magazine.art…scarabin
- I expect this to be more widespread when the cards finally fallscarabin
- "which can get complicated" at the end of the day this is software putting limits on "in the style of" prompting is not some herculean undertakingjonny_quest_lives
- I don’t think “in the style of” is protected by derivative works copyright lawscarabin
- Maybe something featuring mickey mouse would though? I think how it’s being used is considered. Like parody would be allowedscarabin
- "Did you seek consent from living artists or work still under copyright?"jonny_quest_lives
- No. There isn’t really a way to get a hundred million images and know where they’re coming from.jonny_quest_lives
- ummm pay for licensable stock?, Public domain imagery not protected by copyright? there was a more legally sound way to build the datasetsjonny_quest_lives
- this shit will probably fundamentally change the way we all work so they should probably get it right.jonny_quest_lives
- it takes years for these things to go through the courts, it'll be 2030 before any change in law on what constitutes derivative works in AI training datasets,kingsteven
- riskier for commercial entities like MJ. but it's just too far from existing law (largely based on how search engines scrape thumbnails) to predict, given thatkingsteven
- https://techcrunch.c…jonny_quest_lives
- said law wouldn't affect billions of works created from said trained models. the current explosion in AI almost like 'bomb carbon' is to carbon dating post 1952kingsteven
- https://arxiv.org/pd…jonny_quest_lives
- oh nice, i was looking for something like this a few days ago. i've also been cross referencing generated images to the datasets - interesting report.kingsteven
- @kingsteven yeah it all moves so fast... I often use Disney as an IP giant as reference but I actually believe Warner Bros. Lawyers are far more aggressive.jonny_quest_lives
- but that tech crunch article with citated study was super interesting cuz one never realy knows how much of the source is replicated in the outputjonny_quest_lives
- i also think a lot of the IP law and right of publicity law is settled and midjourney will be served cease and desists on certain promptsjonny_quest_lives
- I agree straight up copying and pasting an entire couch or whatever is too far. Maybe put a limit on how much and to what degree things are sampledscarabin
- Find a happy medium somewhere much like artist’s 30% rulescarabin
- How was the dataset built?
It’s just a big scrape of the Internet.
__
So they take what are not from him...Gabriel - yeah gabriel, they are trained on small images, the legal precedent is that search engines can scrape and create thumbnails because it is 'transformative' usagekingsteven
- but legally 'transformative' doesn't mean 'scaled' it actually means unrelated to any aesthetic purpose. so it will inevitably be found to be illegal, in 2040,kingsteven
- considering that we could be in a situation very soon where more AI art has been created and online than human art its fair to say its unstoppable.kingsteven
- ie. even if it was found illegal to train models from LAION-5B dataset at some point. that law wouldn't apply to images generated from the model...kingsteven
- it would be impossible to prosecute an end user or order a mass takedown of images, (unless blatent infringement of IP) so morality is everything RNkingsteven
- I don't get as deep as kingsteven... I think it's easier for companies to go after midjourney for IP infringementjonny_quest_lives
- They have a commercial product that currently allows users to output Marvel characters and noodle it and some might fall in fair use but some doesn'tjonny_quest_lives
- At the bare minimum they could do this:jonny_quest_lives
- "In response to a complaint we received under the US Digital Millennium Copyright Act, we have removed the ability to render for the requested prompt"jonny_quest_lives
- Perhaps Adobe is currently scraping/training their AI of all users storing stuff in the creative cloud at this moment. It's AI Armageddon at this point.jonny_quest_lives
- Month to month subscription and access to users designs/builds... more incentive for AD'S not to name layers.jonny_quest_lives
- companies like midjourney are designed to take that heat... you can't simply issue a cease and desist with no precedentkingsteven
- to find a software company or AI liable it would require a rewrite of not only copyright and trademark law (in every country of operation) but challenge basekingsteven
- legal concepts such as 'originality' it's huge. the more obvious solution would be to prove that prompts constitute originality (skill or labour of an artistickingsteven
- kind) and shift responsibility to the user. with the copyright holders own AI enforcing takedowns on an image-by-image basis.kingsteven
- as far as i'm aware the only realistic legal challenge is what constitutes fair use in training data (mentioned above) that wouldn't affect existing models orkingsteven
- works, and as holtz mentions may not be how AI is trained by the time that is resolved. the first step is the ruling in the andy warhol/ prince artworks casekingsteven
- thats already been several years through the courts. tbh i can't see anything like you propose in the original post happening as things standkingsteven
- was it you saying that this wouldn't happen with music? well you may get what you wish for unfortunately when AI starts enforcing take downs of human artistskingsteven
- - copyright struck for basing their style on Jean Giraud :-s ala Robin Thicke/ Marvin Gaye. Could get real messy, real quick.kingsteven
- To me i think the partisan good/bad arguments are harmful to everyone it's possible to use tools trained on public datasets morally.kingsteven
- Midjourney at the end of the day is a commercial company regardless of their posturing. their product is just software. at some point there will be the legaljonny_quest_lives
- challenge from say Warner Bros. as they are pretty litigious. along the lines of hold up too many near perfect SuperMan illustrations it's at that pointjonny_quest_lives
- that IP and Copyright are largely settled. Midjourney will say well the user enters the prompts it's on them. WB will say disable the promptsjonny_quest_lives
- why are there no safeguards in your commercial software? the messiness of the music example is just as messy as the text to image side of things.jonny_quest_lives
- in the style of "insert artist name" is extremely problematic in both music AI generation as well as text to image AI in my opinionjonny_quest_lives
- the sad state of affairs is it's easier for WB/Disney to hit Midjourney with the "your AI is generating renders of our copyrighted material and charging for itjonny_quest_lives
- WB/Disney will rightfully argue regardless of how it's trained its just a bespoke xerox machine churning out variations of our IP and they charge for itjonny_quest_lives
- Buzz Lightyear on the moon any human artist can paint... Stable Diffusion allows a global user base to render out an Ai generated version based on promptsjonny_quest_lives
- and fine tune it until they are happy. A human artist depending on popularity may or may not ever get hit with a Cease and Desist but legally they would needjonny_quest_lives
- license agreement to sell the work commercially so I would think Midjourney would need similar licensing for WB/ Disney to provide outputs to it's userbasejonny_quest_lives
- if what they're doing isn't illegal under current copyright law MJ aren't financially liable so why would they comply?kingsteven
- from warners perspective using AI for copyright enforcement is the more profitable endeavour and (as i was trying to illustrate with the music example)kingsteven
- going to be harmful to non-AI art. if trained images cant be recreated pixel perfectly from the model the grey area on fair use of images under current lawkingsteven
- for use in training is covered by opt-outs.kingsteven
- ok so here's an example cuz i am struggling here. say a user prompts a batman image and it's in the style of say Alex Ross. and it's beautiful and the userjonny_quest_lives
- is happy. Does nothing with it but Midjourney took a few for rendering the output and they have no license agreement with WB are you saying WB can't pursue ajonny_quest_lives
- **took a feejonny_quest_lives
- are you saying WB can't pursue a Cease and Desist against Midjourney for profiting off of their IP?jonny_quest_lives
- It's sort of the bootleg T-shirt model they may not be 1:1 they also may be original AI Renders but the render is the product and WB hasn't granted Midjourneyjonny_quest_lives
- any rights for their AI to output WB Intellectual Property. It's moved so fast the lawyers are preparing strategy i would assume.jonny_quest_lives
- unauthorized reproduction diminishes their brand/ip and granted it may be a minority subset of users at this point.jonny_quest_lives
- https://www.tiktok.c…jonny_quest_lives
- really sucks that i sound like my dad now too... ffs. maybe i should embrace the whirlwindjonny_quest_lives
- https://www.tiktok.c…jonny_quest_lives
- https://www.artstati…kingsteven
- a lot of the confusion would seem to be that folks think AI has intelligence or sentience like how that tiktok "asks midjourney to"kingsteven
- it's generating images from prompts and noise, there is no law that attributes authorship to software and a judgement that did implicate MJ wouldkingsteven
- challenge other parts of copyright law (such as attribution of copyright to the lifetime of an author), then that judgement would conflict with other countries,kingsteven
- so i'm sure companies like Warner Bros are being advised not to pursue rush judgments because it's a huge risk to their income for no reward.kingsteven
- hypothetically if Warner got a judgment against MJ and MJ disabled prompts, MJ could have right to sue them further down the line.kingsteven
- for example, in the UK there is law that covers generative works that could be interpreted to attribute authorship to the person that creates the promptkingsteven
- but in the US a conflicting judgment would be that gardens (yes gardens) are not subject to copyright because they are cultivated, not authored...kingsteven
- that's just one example, but as i understand it there is no way currently that MJ could be held responsible without challenging core premises of copyright lawkingsteven
- neverscared1
Artificial Aesthetics: A Critical Guide to AI, Media and Design
http://manovich.net/index.php/pr…
Assume you're a designer, an architect, a photographer, a videographer, a curator, an art historian, a musician, a writer, an artist, or any other creative professional or student. Perhaps you're a digital content creator who works across multiple platforms. Alternatively, you could be an art historian, curator, or museum professional.
You may be wondering how AI will affect your professional area in general and your work and career. Our book offers intellectual tools to help us better see the future of creative fields. These tools come from philosophy of art, experimental psychology, media theory, digital culture studies and data science. The book is the first to combine these different perspectives together.
We started the work on the book in summer 2019, exchanging numerous messages, commenting on each other ideas, and sharing drafts of sections. The final book is a result of this process. Although each chapter is written by one author, it reflects the discussions we had over 27 months.
The book is released one chapter at a time on manovich.net and gc-cuny.academia.edu/LevManovich...
- I have seen the bottom fall out of 2-3 of these industries in my lifetime. I don't fear AI I just need it legally cleared before I can touch it or consider itjonny_quest_lives
- In a professional workflow.jonny_quest_lives
- That being said I don't want to feed the beast or be a paying guinea pig either. Nothing I do is magical... I am largely paid to endure the 7thjonny_quest_lives
- Ring of hell of yearly revisions slow marching to a release date that may or may not get pushed. Only to have the marketing teams change directionsjonny_quest_lives
- at the slightest hint of bad press or PR catastrophe.jonny_quest_lives
- It keeps the lights on and the kiddos fed. Grateful for that but I have dreams of one day becoming a Windmill technician. So it’s not too late to follow one's djonny_quest_lives
- Dreams.jonny_quest_lives
- https://youtu.be/g1t…jonny_quest_lives
- sted1
fancy headline, worth the read
"Google might be hitting the panic button to protect Search from ChatGPT"
- scarabin1
Palette.fm colorizes black and white photos
- neverscared-2
“Artificial intelligence” is maybe the wrong way to describe what happened here. This was an art heist.
- Why the DVs?! This is true.grafician
- ethiics dont count... everbody is fucking everyone over more and more in the coming years .. it will be a super clusterfuck...neverscared
- Talking to several friends and this is fairly anecdotal... Art Director's excited about Ai tend to have just a graphic design only background.jonny_quest_lives
- Versus illustrators/photogr... who transitioned into graphic design and became Art Directorsjonny_quest_lives
- 3d digital designer too i would speculate...neverscared
- scarabin3
- Just saw a naked picture of santa with the body of a barbie doll. Hell has been unleashed.PhanLo
- My god there’s some imagination in there
https://www.reddit.c…Ianbolton
- NBQ003
Some good useful stuff
- This is why AI could be super useful in my opinion.ApeRobot
- Has been usefil for game developers for years. Game behaviors, npc's, swarm mechanics.. It's just how much you let to the technology take over your process.jonny_quest_lives
- If this leads to a generation of lazy AI generated art/graphic design we are fucked. Not optimistic based on past human behaviors.jonny_quest_lives
- aaaand at 6:55ish the infamous "Greg Rutkowski" prompt appears (－‸ლ)jonny_quest_lives
- sted0
While recent work on text-conditional 3D object generation has shown promising results, the state-of-the-art methods typically require multiple GPU-hours to produce a single sample. This is in stark contrast to state-of-the-art generative image models, which produce samples in a number of seconds or minutes. In this paper, we explore an alternative method for 3D object generation which produces 3D models in only 1-2 minutes on a single GPU. Our method first generates a single synthetic view using a text-to-image diffusion model, and then produces a 3D point cloud using a second diffusion model which conditions on the generated image. While our method still falls short of the state-of-the-art in terms of sample quality, it is one to two orders of magnitude faster to sample from, offering a practical trade-off for some use cases. We release our pre-trained point cloud diffusion models, as well as evaluation code and models, at this https URL.
- sted-1
https://github.com/AbdullahAlfar…
With Auto-Photoshop-StableDiffusion-P... you can directly use the capabilities of Automatic1111 Stable Diffusion in Photoshop without switching between programs. This allows you to easily use Stable Diffusion AI in a familiar environment. You can edit your Stable Diffusion image with all your favorite tools and save it right in Photoshop.
- sted0
- NBQ00-1
- Krassy3
- brilliant. but it's all random it seems. I copied the same prompt to test and got this https://playgroundai…_niko
- so I imagine it's a lot of trial and doors and tweaking, or do they upload a drawing as a bas and hav the ai make it better?_niko
- It’s like a fruit machine. But a base photo could provide good resultsPhanLo
- yeah playing around with some base drawings I've done, super cool and eliminates a lot of the randomness_niko
