Quote from: MatHayward on July 21, 2023, 07:12
Hi Everyone,

I would like to share two updates regarding the submission of generative AI images to Adobe Stock. We have updated the contributor portal and added an additional checkbox that eliminates the need for you to attach a property release for generative AI images featuring people or property that are fictional.

In addition, you can now submit photo-realistic generative AI images as asset type photos, i.e., content that looks as though it could have been created by a camera, as long as you check the box that it was created with generative AI tools. All other generative AI content should continue to be submitted as asset type illustrations.

You can read the details on all the updates in our learn and support page here:
https://helpx.adobe.com/stock/contributor/help/generative-ai-content.html

Let me know if you have any questions.

Thank you,

Mat Hayward

Thank you very much Mat!
Clear and easier now, and (crossing my fingers) quicker 

I'm following the blog ( https://www.technollama.co.uk/) of  Dr Andrés Guadamuz expert of  Intellectual Property Law at the University of Sussex (also quoted in a post by Jo Ann)
His articles cover topics on copyright, infringement, artificial intelligence and machine learning.

Here a draft  : A Scanner Darkly: Copyright Infringement in Artificial Intelligence Inputs and Outputs
https://drive.google.com/file/d/19fVgFAOK0X0Mwbp-Z5w5WcaeTO8Cx7Lx/view?pli=1

and the last article Authors sue OpenAI for copyright infringement https://www.technollama.co.uk/authors-sue-openai-for-copyright-infringement

Enjoy (as I do) 

Quote from: cobalt on June 30, 2023, 10:09
@gameover

congrats!

Also trying to do more concepts. But I need a lot more experience to get them right.

Just think out of the box, it's easy! 

Quote from: alijaber on June 29, 2023, 23:25

Quote from: alijaber on June 09, 2023, 19:34
I've encouraged my 17 years old son to create photos using Midjourney for AS. He doesn't know anything about photography. He started uploading images to my AS account 3 months ago, here are the number of sales (remember he's novice):
April: 27 sales
May: 51 sales
June (till 9): 28 sales
He now has 1500 files online (260 under review) and he's planning to upload 500 - 1000 images per month.
I do videos mainly, but I have had around 300 images online for years, and he already exceeded my number of sales for images.

Updated numbers (really interesting experiment) ...

Congrats   

That's really a good start!

Looks like he found out that the idea behind the images plays a decisive role: once I realized it my sales skyrocketed (this week a single image got me $21.25 and another $26.40)

Quote from: cobalt on June 27, 2023, 13:32
LOVE IT!

Please upload that, it will bring great money

thank you! I took your suggestion 👍

Quote from: DiscreetDuck on June 27, 2023, 14:11

Quote from: gameover on June 27, 2023, 13:58
BETTER?

I feel sorry but... No.
Why do you never check the images that AI generates for you! It might help you understand some pitfalls
Here is a correction:

I hope you did not already upload it to the stock sites! You should correct this before.
There are buyers who learned reading and writing, even for a language other than their mother tongue 

the trolls are not cultured people...

Quote from: gameover on June 27, 2023, 13:40

Quote from: cobalt on June 27, 2023, 13:32
LOVE IT!

Please upload that, it will bring great money

I forgot to add a placard "politeness is obsolete!" - working on it 

BETTER?

Quote from: cobalt on June 27, 2023, 13:32
LOVE IT!

Please upload that, it will bring great money

I forgot to add a placard "politeness is obsolete!" - working on it 

Quote from: DiscreetDuck on June 27, 2023, 11:32

Quote from: cobalt on June 27, 2023, 11:22
Why are you going on a witch hunt against people or agencies that use ai?

I see many people singing the praises of AI...
I only feel free to have my own opinions. An issue for you?
do you live in a dictatorship? or would like to?

(made with AI  )

Quote from: Jo Ann Snover on June 24, 2023, 14:39
Midjourney is adding new features - zoom out. The comments are interesting as well as the article itself

https://arstechnica.com/information-technology/2023/06/stunning-midjourney-update-wows-ai-artists-with-camera-like-feature/

I used it, here my first attempt: very entertaining 



Midjourney 5.2 tells stories!
you can see the sequence here
https://luisafumi-digitalart.com/blog/2023/06/24/midjourney-5-2-tell-stories/

Quote from: Evaristo tenscadisto on June 12, 2023, 19:19

Quote from: gameover on June 09, 2023, 20:08

Quote from: wds on June 09, 2023, 19:52
I'm guessing one important factor is as follows:

  - Many photos uploaded to Stock sites aren't super saleable....i.e. snapshots, nice landscapes....etc.. Generally with stock, photos with people in "interesting" locations (e.g. offices) sell better. For a photographer, that means hiring models with releases etc., and getting those interesting locations. For AI, it is no problem to include "people" as subjects in "interesting lcations". In general, with good prompts, AI can produce saleable images without having to hire actual models or travel to interesting or challenging locations as a photographer would have to do.

Right!
The AI is a powerful tool and another resource at your disposal. Important is the choice of the subjects and the post processing. Anyway some of my AI generated images are performing very well with more than 50 downloads each and others are approaching...

Congrats! 

thank you 

Quote from: wds on June 09, 2023, 19:52
I'm guessing one important factor is as follows:

  - Many photos uploaded to Stock sites aren't super saleable....i.e. snapshots, nice landscapes....etc.. Generally with stock, photos with people in "interesting" locations (e.g. offices) sell better. For a photographer, that means hiring models with releases etc., and getting those interesting locations. For AI, it is no problem to include "people" as subjects in "interesting lcations". In general, with good prompts, AI can produce saleable images without having to hire actual models or travel to interesting or challenging locations as a photographer would have to do.

Right!
The AI is a powerful tool and another resource at your disposal. Important is the choice of the subjects and the post processing. Anyway some of my AI generated images are performing very well with more than 50 downloads each and others are approaching...

Quote from: Evaristo tenscadisto on May 21, 2023, 19:01
Lowls,
What prevented anyone before AI from writing their own ideas in one style or another?  In fact, all artists start by absorbing styles from other artists that they like and influence them to write. Isn't that the normal learning process for humans?

Taking the example of music. How many bands exist and still exists with a Beatles-style sound all over the world? I would say hundreds if not thousands.I remember the first time I heard the band Oasis and I felt a kind of revival of the Beatles sound. Naturally there were particularities such as themes, some writing and type of voice but in essence it reminds me a lot of the Beatles.
Guitarist Van Halen developed the unique style of "Fingertap" on the guitar. Among others Steve Vai or Joe Satriani use this style in their music/songs. If we are going to talk about good BB King then practically most guitarists use the famous "bending" style which makes the guitar seem to cry especially in Blues, Rock or variants...

Jumping to graffiti how many artists "imitate" banksy?
Despite the Stencil technique, the Grunge style of drawing, and the Transformism of images such as the Clown Ronald Macdonalds and Mickey Mouse holding hands with the Napalm Girl are not his, the originality in the portrait subject marks his authorship.

If you look at Andy Warhol's work well then he took pictures that weren't of him and painted over them (i.e. Marilyn Monroe) Warhol created this masterpiece which consists of 50 images of Marilyn using the same publicity photograph from the film titled Niagara.

So nothing stopped any of them from absorbing and learning styles and bringing originality to our world, right?
AI will not kill creativity. In fact, it will do the opposite. It allow you to explore and develop many more styles or combinations of styles than any artist could do in his lifetime. This is where originality and a sense of art are born. So AI can help you to perform a better drawing, painting, writing or re-write your work. Also can bring new things to your creative table that you may include or not. It's really up to you. If you are having trouble in understand if AI can substitute the artist it can not. AI is only an excellent performer. Do not forget who write the prompt.

On the other hand of course you can use AI to just copycat but we don't need AI for that we've been living with this problem for decades and normally they don't go so far without being noticed.

I will not extend my post any longer since i find you were more interest in AI for Artistic vision which is a small water drop in the AI ocean. AI will assist you in everything you do - literally everything! You cannot use fake data/metadata for research. The redundancy, entropy and risk is too high but you can run simulations based on real data/metadata to help you with accuracy for some diagnosis.

It is a great relief to read again a voice of reason in here!  I was afraid all reasonable souls had run away scared by the Bandar-log   

https://www.reddit.com/r/ChatGPT/comments/13morhd/chief_ai_scientist_at_meta/



Reminds me of something...

Quote from: Justanotherphotographer on May 16, 2023, 19:25
The point was to demonstrate that the images are stored in some form which was demonstrated by retrieving them with the right prompt. The relevant point is that the images are, in fact, compressed in a database that is pulled from. I am on mobile so can't  get to the paper right now. Doesn't it conclude that something like 0.03% of the time they were getting recognisable results, so almost 1 in 300? (tough this is really irrelevant, the point is they conclusively demonstrated that original images are stored, how good the app is at covering it up is less relevant).

ETA. Just reread your post. Lol made up stats indeed. You took the total database size and divided it by the 50 images to calculate the probability? Come on, you're a scientist.

Oh, about the stats please teach me! And about the storage, please consider recommending Stable Diffusion not to keep their disks in a mouldy cellar and use the cloud instead 

Quote from: Justanotherphotographer on May 16, 2023, 12:26


Here's the actual paper:

https://arxiv.org/pdf/2301.13188.pdf

When reading bear in mind all the major AI generators now use diffusion models rather than GANs. I described the method before and was told "that's not how AI works" by someone who then proceeded to describe GANs.

The dataset they used to extract images is Stable Diffusion "the largest and most popular open-source diffusion model. This model is an 890 million parameter text-conditioned diffusion model trained on 160 million images."

Here's an extract:
"Examples of the images that we extract from Stable Diffusion v1.4 using random sampling and our membership inference procedure. The top row shows the original images and the bottom row shows our extracted images.
"

Here is what they actually use the smaller data set to investigate (in the second part of the paper):

"The above experiments are visually striking and clearly indicate that memorization is pervasive in large diffusion models—and that data extraction is feasible. But these experiments do not explain why and how these models memorize training data."

So they didn't use a hand picked dataset to prove they could extract the images at all. They used the largest set available. Only subsequently did they use a smaller set to show how they were stored.

I read the paper: while aiming specifically at Stable Diffusion (with images based on the LAION dataset) using a specific algorithm  they managed to "extract" 50 images out of 175 million, and all those 50 images were duplicated at least 100 times in the dataset. In order to retrieve the images they had to use as a prompt a string siphoned from the LAION dataset itself. Speaking about doctored stats...

One would conclude that, unless you're specifically hunting for a scandal, the probability of getting a "tainted" image from Stable Diffusion is 1 over 3,500,000.
An average human life lasts (with a bit of luck) about 29,000 days, hence - statistically speaking - you'd have to create with Stable Diffusion 1206 images a day (starting on the day of your birth) before getting one.  Roll up your sleeves... 

Quote from: RalfLiebhold on May 08, 2023, 20:42

Quote from: gameover on March 25, 2023, 11:34
A nice surprise this morning: 4 sales for a $694.22 total (gross), 40% mine. First time it happens all in one day (Champagne !!!) 
SS would give me $0.40 and keep the rest

Congratulations  , I would prefer this to champagne:

https://www.vladi-private-islands.de/en/

mmm... interesting... maybe a second thought after the next 2,000,000 photos sold 

Quote from: dragonblade on May 08, 2023, 13:15

Quote from: gameover on March 25, 2023, 11:34
A nice surprise this morning: 4 sales for a $694.22 total (gross), 40% mine. First time it happens all in one day (Champagne !!!) 

Woah, that is awesome!

 

Quote from: X9D7CE84A2B5Y on May 03, 2023, 11:07
What are content review times normally like right now on Alamy?

Mostly immediately after upload (I bet they use AI for pre-screening), else in 24 hours.

Quote from: cobalt on May 07, 2023, 14:15
Here is a 34 year old writer who lost his income because of Chatgpt

https://www.reddit.com/r/ChatGPT/comments/139o1q6/lost_all_my_content_writing_contracts_feeling/

I've just read it. One particular sentence struck me, and I couldn't resist commenting (on site) :

"For me, writing is like taking a sh*t: I don't have a choice."

We all do that, we just don't sell it.

Take it easy, plumber is quite a dignified profession. Perhaps more than selling words...

However I suspect that's just a prank: check his grammar, his spelling, his style, and then ask yourselves if anyone could really make a living out of that.

Dismiss in the class action lawsuit against StabilityAI, Deviantart, and Midjourney

https://twitter.com/technollama/status/1648981345924685824?cn=ZmxleGlibGVfcmVjcw%3D%3D&refsrc=email

Dall e 2 will make us all redundant?
To expand this debate a little further, my last blog post



https://luisafumi-digitalart.com/blog/2023/04/19/to-be-or-not-to-be-lets-ask-chatgpt/

The Ai-generated  images sell pretty well,  but not by themselves.
My best seller has been download at Adobe 52 times - but as all my AI images it went through a careful postprocessing in Photoshop and other dedicated software. Not only to remove the obvious errors (too many fingers, missing eyes etc.) but also to give them a particular light and style - they must reflect the way I feel, not MidJourney.
Also devising the right prompt is no kid's play: the AI is still too primitive to guess accurately enough what one has in mind upon a few words.
Last but definitely not least, the human creativity lies in the idea behind the prompt - mediocre prompt, mediocre outcomes. I've been waiting long for someone to mention this here, but so far in vain...

Quote from: ribtoks on April 05, 2023, 10:02
Hey folks

A bit overdue, but finally I got to write a long blogpost about everything you can do today with generative AI in the context, useful for microstock contributors.

One spoiler quote:

QuoteWhen was the last time you got a 40% productivity boost from using any tool the money can buy?

(yes, you might not get replaced during the next couple of years)

Let me know what do you think about this: https://xpiksapp.com/blog/generative-ai-microstock/

Cheers

As promised, I've read your article thoroughly and attentively. It's a great article, informative and unbiased, and I totally agree with your conclusion "you will not be replaced by AI, but you will be replaced by people who use AI".

The sad gist of it is, only those able to stimulate an AI with some creativity will have a possibility. The others will have no other option than joining the naysayers or looking for another way out (another niche perhaps?).

The problem is anything but new: since the dawn of time every new technology has left some people jobless by doing their job faster, better and cheaper. And every time the Luddites succumbed while the ones who were quick enough to jump on the new horse prospered.

Quote from: Uncle Pete on April 05, 2023, 18:56

Quote from: gameover on April 05, 2023, 12:11

Quote from: ribtoks on April 05, 2023, 10:02
Hey folks

A bit overdue, but finally I got to write a long blogpost about everything you can do today with generative AI in the context, useful for microstock contributors.

One spoiler quote:

QuoteWhen was the last time you got a 40% productivity boost from using any tool the money can buy?

(yes, you might not get replaced during the next couple of years)


Let me know what do you think about this: https://xpiksapp.com/blog/generative-ai-microstock/

Cheers

Thank you very much for your quite interesting post!
The first impression is of an exhaustive and very well balanced article - now give me time to read it and appreciate it thoroughly

But I'm afraid that to some people here it will be like preaching to the windmills: they are set on mourning and mourning they will till their last breath

Yes you predicted the replies below yours, before they were posted.  S O S and the first two words are Same Old 💩



Hey Yes to that conclusion: "In the end, you will not be replaced by AI, but you will be replaced by people who use AI."

Dall-E2 says: "The server is currently overloaded with other requests. Sorry about that! You can retry your request, or contact us through our help center at help.openai.com if the error persists."

S.O.S.  ... --- ...
You made my day 🫀

As a matter of fact my crystal ball is pretty old and foggy and runs on good old Win7, but to predict some people it is more than enough...



(the fingers are okay!  good old Midjourney)