I don't believe it,I don't think it's useful to train an AI on the basis of other AI images,and then I also don't believe it because they say that Adobe wanted to train Firefly in an economical way,it doesn't seem like it to me,given that Adobe even made an effort in searching for content through paid missions for contributors.

if this were true,then why bother even with paid missions?

then I don't know if I read everything correctly,but I don't think any official Adobe representative has ever said that they used other AI images to train Firefly.

Then I don't understand this sentence:"Some of those artists submitted images generated by Midjourney and other rival AI,and were compensated by Adobe for their input,according to Bloomberg's report".

who?when?

but I'm more than sure that Firefly was trained not only with the Adobe Stock library,but they certainly didn't use other AI content and I'm more than sure that they always paid the rights of this content.

Don't believe everything you read!

I have said this (on Discord) from the moment when I found out that AI-only contributors were paid the Firefly bonus. It makes a mockery of Adobe Stock claiming to be ethical. They pass it off by saying that any AI content submitted and approved to their site has been designated by the contributor as non-infringing. But most of that AI content has Getty Images and Shutterstock and other stock agencies' watermarked images in its dataset.

@InJuisticeForAll On the Discord server Mat Hayward has said they used Ai images to train Firefly.

I have said this (on Discord) from the moment when I found out that AI-only contributors were paid the Firefly bonus. It makes a mockery of Adobe Stock claiming to be ethical. They pass it off by saying that any AI content submitted and approved to their site has been designated by the contributor as non-infringing. But most of that AI content has Getty Images and Shutterstock and other stock agencies' watermarked images in its dataset.

@InJuisticeForAll On the Discord server Mat Hayward has said they used Ai images to train Firefly.

ok,if Mat said it I can start to believe it,however,I have only seen the contents with watermarks and various signatures on Midjourney.

I have never seen anything in the content generated by Firefly that could be attributable to content taken from other agencies,such as watermarks or signatures.

AI-only contributors were paid the Firefly bonus?this is new,have you been able to personally verify this?

as I said,I find it extremely difficult to think that an AI can be trained on the already generated renderings of other AIs,but ok,if Mat said it I can believe it!

Quote from: Injustice for all on April 20, 2024, 09:23

as I said,I find it extremely difficult to think that an AI can be trained on the already generated renderings of other AIs,but ok,if Mat said it I can believe it!

Training AI with AI is a slippery slope kind of thing.

Adobe has said they used up to 5% AI images, which were individual reviewed and vetted, to train the Firefly lab. There's still debate on this, which I'd agree, that using AI to train AI is wrong. Not ethically wrong, but scientifically and system integrity kind of wrong.

Quote from: DiscreetDuck on April 20, 2024, 08:40

I(we) would NEVER have sold rights to use my(our) photograhies to feed machine learning. Adobe forced me(us) by giving money, to infringe my(our) copyrights on MY(our) own images. It seems to me that no choice was offered to refuse this money. In this way, they can suggest a mutual understanding, in possible legal procedures, and claim to act ethically.

I'm not going to say I support how the agencies did this, but I have to answer, that you signed the contract, that allowed this use and you could have refused the money, but they still had the right to use your images. You made that choice when you agreed to the contract.

"I(we) would NEVER have sold rights to use my(our) photograhies to feed machine learning."

You Did.

If you think that's wrong and disagree, you and all those silent people, who aren't here according to you, should get together and file a claim against Adobe. Class action suit if there is any attorney who will see your side of the arguments, after reading the contract that you signed, when you offered to be a contributor to Adobe or Fotolia.

There you are. You signed a contract, now you disagree. Take it up with the courts.

Actually, there is no clause that say they can steal your images to train competing software.

Also - some of the agencies (i.e., shutterstock) refused to 'take back' the money. It was "forced", dishonest.

I already critisized this when it was happening, but not all that many people seemed to care?

https://www.microstockgroup.com/fotolia-com/announcing-bonus-payment-for-adobe-firefly-training/

Quote from: MatHayward on September 13, 2023, 09:09

Quote from: Her Ugliness on September 13, 2023, 08:47

Mat, do I understand this correctly, that the payment is solely based on portfolio sice and sales, not on whether the images were actully used to train your AI? Because if that's the case  that basically means that someone who has not done a single real photo or illustration in all his life, but has only created tons of AI content based off other people's hard work also got the "compensation". So they basically got PAID for generating AI images based off the work of people who never got paid for it.
That's a far cry from Adobe's statement that you were "developing generative AI responsibly, with creators at the center."

The initial bonus is based on the all-time total number of approved images you submitted to Adobe Stock and the number of licenses that those images generated in the 12-month period between June 3rd, 2022 to June 2nd, 2023. The bonus is weighted towards license.

-Mat Hayward

Yes, I can read. I understood that very well. I just wanted to clarify whether that REALLY means that you also compensated people who only submit AI images, so never had any images to offer for your training in the first place. Because that seemed insane. But apparently that's the case as you just copied the info from the mail Adobe sent out without adressing the actual issue.
Very "fair". 
So far, of all the agencies that use their database to create AI content, Adobe seems actually to be the most unethical one! No opt-out option AND giving money that was meant as compensation for having our real photos and illustrations used to train your AI to people who only use our content to generate AI images. I am sure these tons of new "contributors" that only submit AI content are overjoyed from now profitting even more from other peoples' work. But Adobe seems to care more about these pople than real photographers and illustrators anyways.

Adobe does not care whether their AI is really ethical. They just want to put that label on it to tell their customers it was ethical and safe to use. It's nothing more than a marketing stategy to make more money and nothing is ethical about it. It's  like green-washing, but for AI -  Ethical-washing. 

I beg to differ on this,the way Firefly was trained in my opinion is always the most ethical,even if they used AI images already in Adobe's collection.

Midjourney literally stole content for its own training,without even bothering to remove logos,signatures,watermarks,not paying anything to anyone,and asking substantial sums for subscriptions.

Firefly paid for all the content used,they didn't steal anything from anyone,and they didn't upload content with logos,signatures and watermarks,also leaving free access to the platform for months.

No. Just absolutely NO.

You accuse of midjourney stealing content. But then Adobe knowingly used that stolen content from Midjourney images to train their AI. What's the differene? NO. We did not get paid for that. MY images were used to train Midjourney. Adobe used Midjourney images to train their AI. Where in that chain did Adobe pay me to use Midjourney images based on my stolen content to train their AI?!

And you pretend as if logos, signatures and watermarks were all that matters? As if using images to train an AI without the creator's permission was somehow okay as long as the stolen images had no logos or watermarks?

All Adobe did was let Midjourney do the dirty work for them so they can pretend their own hands were clean.

Actually, there is no clause that say they can steal your images to train competing software.

All the agencies have super wide terms, aka weasel words, which they can subsequently use in various ways we, and probably even they didn't think of at the time.
Adobe clause 1:
"Licenses for Adobe.
You grant us a non-exclusive, worldwide, perpetual, fully-paid*, and royalty-free license to use,
reproduce, publicly display, publicly perform, distribute, index, translate, and modify the Work for the purposes of
operating the Website; presenting, distributing, marketing, promoting, and licensing the Work to users; developing new
features and services; archiving the Work; and protecting the Work. "

Of course you didn't think that meant training AI, but the permission is there for them to do so.
Notice also that sneaky 'perpetual', which covers them even if the contributor pulls their port.

* "fully-paid"?   I have no idea!

Did Adobe steal something?No
Did Midjourney steal something?Yes

What about the notion of receiving stolen goods?
And make it the source (even partial) of a business?

Quote from: DiscreetDuck on April 21, 2024, 06:51

Quote from: Injustice for all on April 21, 2024, 05:15

Did Adobe steal something?No
Did Midjourney steal something?Yes

What about the notion of receiving stolen goods?
And make it the source (even partial) of a business?

so you assume that content generated by AI is stolen content?

so if you have AI content in your portfolio you are a thief too!

whereas if you don't have AI content in your portfolio,you are consistent with your ideas,and I respect that,but I don't agree with it!

It's a little bit of a "game" some people play.

Say someone steals a vehicle then resells it to you, and you didn't know it was stolen. Did you "steal" it? No - you paid for the vehicle - thinking it was good. However - the fact is, it is still a stolen vehicle, regardless of whether the person stole it sold it to you, or you now "own" a stolen vehicle.

Now question is slightly different. Say someone steals a vehicle then resells it to you, and you DO know it was stolen. Did you still "steal" it? I'd say it is a debatable point/gray area - but ultimately/"technically" no - because you paid for the vehicle, and you didn't do the actual 'theft' - so you didn't "steal" it - but you did know it was stolen. But fact is, in this case - it is still a stolen vehicle, regardless of whether the person stole it sold it to you, or you now "own" a stolen vehicle.

Is it 'right' (ethical) in ethical case? I'd say no - however... The large companies (i.e., 'midjourney') have set up an environment in which to "compete" effectively - it almost becomes necessary to use their "service" (which deals in pure stolen goods, the ceo according to google is also apparently jewsih- which is significant in that that seems to be the modus operandi of that culture and it's just how their businesses tend to operate, as the backers also are, the major investors, you need to understand the culture to understand how their businesses operate - but entirely different topic)...

So... then the question becomes...
a) If an environment is set up in which stealing is encouraged...
b) Sellers, in order to compete - feel pressure to buy/sell stolen goods... (And I think potentially a lot of creators don't fully realize that the "ai" systems are simply sophisticated theft, and might actually think they are using a 'magical' tool that can 'think', which of course it is not)...
c) Buyers willingly (either knowingly or unknowingly) participate in purchasing stolen goods... (i.e., it is possible and likely many buyers don't realize "ai" is simply sophisticated theft, and are actually under the illusion that it is some "thinking" machine, which of course it is not)...

To further add to that though... there is a small 'creative aspect' to the goods that the sellers have, albeit small - in that they do actually create a "prompt" (which is a 'new' item), and 'package' the goods (i.e., keywording, titling, etc)... in which case it is debatable - the goods that the "sellers" have - since "some" creative input was required - so it is not "direct" theft - is it still a "stolen good"? It is a "derivative" of stolen goods (i.e., it was 'based' off of stolen goods) - but does now the new "product" (which did require unique/creative input to create) - make it "not stolen"?

But then I suppose one could argue... if a vehicle was stolen, and the VIN on the car was changed, the car repainted - would one say that was still a "stolen vehicle" or a "new vehicle"...?

But it's a strange environment... when you have (primilary) east indians downloading complete portfolios (outright theft), making a new account to sell goods just to play whack-a-mole... jewhsi backed companies promoting theft-based environments (i.e., chatgpt jewsh/sam altman, midjourney jewsh/david holz) + backers that appear (through various shell organizations/etc) ultimately to be backed by blackrock/vanguard/etc (alsojewish, i.e., larry fink/blackrock)...

So how does one remain honest/ethical AND competitive, when dealing with people employing these dishonest tactics?

One super EASY solution though is the following...

Since companies like midjourney/chatgpt are based off of theft... Put pressure on them to make perpetual micropayments to the contributors whose works were stolen... Of course - they don't "want" that  - because they are extremely greedy and have a pyschotic sociopathic desire for "control" - but really doesn't matter what they "want" - because this would make it right. Also - with that system make it so contributors can choose what % micropayments they want, and if they don't want them, to opt out of the system. Despite what any wannabe-technie might say - it is EXTREMELY EASY to implement... "They" (the companies) have the meta data of where they stole the assets from, and it would be very easy (progrmatically) to set up a system to compensate contributors for their stolen works - and pay them in perpetuity (i.e., monthly for the rest of their lives) every time one of their works were referenced in a new "ai image" or "ai video"... It is also possible to make RETROACTIVE payments to all people's whose works were stolen.

At the same time... create awareness campaigns (social media/blogging/etc) to raise general public awareness that midjourney/chatgpt/etc are based off of stolen goods... regardless of whether they play the game saying "oh they simply got images from a RESEARCH company, so THEY didn't know it was stolen! <wink wink>". yeah, <cough cough bullsh_t>)... ultimately, they knowingly created tools based of massive scraping & theft, and want to re-package the stolen goods as their own, and are trying to create a system in which for a person to effectively compete need to use "their" tools (at which point they could grant/deny people access on a whim - they are pyschotic sociopaths and simply want "control", in addition to greed). That there is nothing "intelligent" about the current "AI" ("artificial intelligence"), and that it is simply sophisticated theft & pattern re-arrangment... So raise those types of awareness campaigns to help put pressure on them to do the right thing.

THEN... the tools are based off of ethically purchased/licensed goods, the derivative works become ethnically purchased/licensed, and a good situation all around...

That is one big solution in order to resolve the dilemma.

@DiscreetDuck

you have all my respect,because you are consistent with your ideas.

unlike others who accuse Adobe of stealing,or unethical behavior because they used AI content to train Firefly,but then have AI content in their portfolios.

in my opinion,Midjourney will have maybe,for me 60% no and 40% yes,legal consequences in the future,Adobe will not.

because stealing content to train an AI is one thing,while using already generated AI content to train an AI is another.

in my opinion,as I have already said,Firefly is without a doubt the AI ​​that has been built in a more ethical way than other better-known AIs,such as Midjourney.

Does pushing a "generate" button makes you copyright owner?

Yes

US Copyright Office: "Your work is under copyright protection the moment it is created and fixed in a tangible form that it is perceptible either directly or with the aid of a machine or device."

Can you register a copyright for protection of an AI image?

No. It must be created by a human.