pancakes

MicrostockGroup Sponsors


Author Topic: I will never use AI  (Read 54540 times)

0 Members and 1 Guest are viewing this topic.

« Reply #50 on: November 29, 2023, 03:18 »
+3
Take Adobe Firefly: "Generate images from a detailed text description." This is the promise of Adobe.

If you write Tiger you get a tiger, if you write Pope you get the image of the Pope.

There must have been a photographer who took a photo of the tiger, and another photographer who took the photo of the Pope." Both photographer go empty and Adobe makes the deal.

Am I wrong?


« Reply #51 on: November 29, 2023, 07:07 »
+2
eta

...then I am just adjusting the clothes a bit, changing out the sky, removing the glasses, changing the hairstyle...


Yes, I have also exchanged skies. For many years now. But it was always a sky that I photographed myself. I've never used a sky from one of your photos. That wasn't allowed either.
And that's exactly what the AI providers get around. Without paying a single cent for it. Why are you not allowed to use elements from other people's images, but AI is? Can you explain why?...

one more time -- i don't know your specific knowledge, but, in general, complaints about being victimized show an underlying ignorance of how these models work.  -- AI does not use ANY elements from images when creating new images - in an entirely separate process it trains on billions of images to create its dataset.  when creating a new image (which may take millions of steps) it no longer has access to the original hundreds of millions of images it  used in training. there are many descriptions, of varying detail, on how this actually works - posted frequently here & available online, so there's really no excuse for continuing to promote this false idea.

the completely separate argument is whether there should be any payment for images to be used in training, but no one has been able to show that pieces of their image shows up in a new creation

To say it right away: I don't have any specific knowledge because I haven't experimented with AI yet. In this respect, I agree with you.

So I can only try to draw conclusions from what I can see.

Maybe I really don't understand the working principle of AI software.

But, if it were as you say, that AI does NOT use ANY elements from existing images, how can it be explained that, for example
- the Apple logo
- the Apple mouse
- the iMac
- the keyboard
- the Mercedes star
and so many other elements can be seen unchanged from the original in the images? Then why doesn't the AI "design" a new Apple or Mercedes logo, a different foot of the iMac, a new mouse etc.? I see here exactly the design features of Apple (material/color, radii, shapes, etc.).

And, if it is as seen here, who can rule out that elements from your and my pictures appear 1:1 in other pictures.

If you have a link that helps me to understand this, I would be grateful.

« Reply #52 on: November 29, 2023, 10:09 »
0
The topic here is: "I will never use AI".
To all the pro-AIs who sell their faith and prostration on this thread: Why not open a topic "I pray I will get eternal AI" for the simple sake of consistency and intellectual respect?  ???

[Edit]:I know that most AI critics tend to leave this forum, and that's unfortunately really understandable.


How to be.... let's say discreet and approach the topic " i never use AI" without actually sharing what people think about it?
Isn't that the idea of participate in the topic of a forum?

Even if it is a different point of view it does not make sense to me not to be included.
There's a well-known saying that fits perfectly here and I couldn't agree more: "It is from discussion that light is born"

People are exposed to different cultures and levels of technology so opinions regarding topics must be open and diverse. Otherwise we fall in the wrong conviction that things are just like some topic title. Some people think that earth is flat but if a topic arise with that title i will gladly share a different opinion about it. Not saying that i am right or wrong but sharing and listening different point of views empowers us to have a better and wide view of how things really are.

« Reply #53 on: November 29, 2023, 10:28 »
0
We've been discussing the topic of AI here for years. Sharing information, knowledge, experiences about what is AI and how could affect us...

I do think we need much more AI regulation and legislators are just trying make up for lost ground. They should consult more with academics and perhaps establish a security committee for new technologies, where the general body is represented by individuals with stronger ties to academic research instead of companies interests. Currently, we only have powerful lobbies that exclusively represent the business side, exerting influence on the legislation of laws, and this is not ideal. I don't mind they exist since they are the engine of economy but there is also a need for a perspective more driven by the social component to balance opinions within decision-making bodies.

I am not against of getting paid to feed AI learning but it cannot be in a royalty perspective for 2 reasons: once used to learn there is no need to re-use it, secondly large amount of data and metadata is required so algorithm cam learn properly.

For the second, even a budget of 1$ per data/metadata is a fantasy for the short budget of the research and developments projects. I mean we are talking about something that will not be even be distributed only used to learn. In this sense i am not talking about midjourney, dall-e or Stable diffusion. I am talking about thousands of projects that require real image data to learn and probably 5 cent is too much.

Furthermore "I never use AI" topic will mean that in future someone will not use smartphone, computers, Tv, internet, electric cars, hospitals, etc. It's a complete return to old "Zoe" but not really a return to innocence. Well...In some way it can be achievable if someone isolates from society.

« Last Edit: November 29, 2023, 10:32 by Evaristo tenscadisto »

« Reply #54 on: November 29, 2023, 10:51 »
0
Take Adobe Firefly: "Generate images from a detailed text description." This is the promise of Adobe.

If you write Tiger you get a tiger, if you write Pope you get the image of the Pope.

There must have been a photographer who took a photo of the tiger, and another photographer who took the photo of the Pope." Both photographer go empty and Adobe makes the deal.

Am I wrong?

Adobe firefly "Trained on Adobe Stock images, openly licensed content, and public domain content, Firefly is designed to be safe for commercial use. To ensure that creators can benefit from generative AI, we've developed a compensation model for Adobe Stock contributors whose content is used in the dataset to retrain Firefly models" link: https://www.adobe.com/sensei/generative-ai/firefly.html

I am curious and If you write your name... does a photo of you appear? And if you upload an image of you instead of text what you think will happen? Have you tried? ;)


Uncle Pete

  • Great Place by a Great Lake - My Home Port
« Reply #55 on: November 29, 2023, 12:26 »
+1
Furthermore "I never use AI" topic will mean that in future someone will not use smartphone, computers, Tv, internet, electric cars, hospitals, etc. It's a complete return to old "Zoe" but not really a return to innocence. Well...In some way it can be achievable if someone isolates from society.

Years ago a good friend who works for the post office said "I will never use email" His reason was, it was taking away from the USPS income as people didn't need to send letters anymore. His email is hisname2000@, because that's the year he decided he needed to have and use email. Didn't take him long to change his mind.

I don't find any need to use AI, to generate images. Just like some people live in a city with good public transportation or ride a bicycle and don't own a car. Some others may find a use or need, for personal transportation, maybe for work? Where I live and it's 4 miles to the grocery store, and two to a gas station. Many other tings I have to go 12 miles. Individuals have cars. It's 28 miles to work. I have a distinct need.

I can agree that by choice, if someone decides they never want to use AI and don't need it, that's their personal choice and depends on what they do. I'm in the middle, I use AI for fun and for free. Other people here, obviously are more serious and use it for work and income. To each their own.  😎

« Reply #56 on: November 29, 2023, 12:30 »
0
eta

...then I am just adjusting the clothes a bit, changing out the sky, removing the glasses, changing the hairstyle...


Yes, I have also exchanged skies. For many years now. But it was always a sky that I photographed myself. I've never used a sky from one of your photos. That wasn't allowed either.
And that's exactly what the AI providers get around. Without paying a single cent for it. Why are you not allowed to use elements from other people's images, but AI is? Can you explain why?...

one more time -- i don't know your specific knowledge, but, in general, complaints about being victimized show an underlying ignorance of how these models work.  -- AI does not use ANY elements from images when creating new images - in an entirely separate process it trains on billions of images to create its dataset.  when creating a new image (which may take millions of steps) it no longer has access to the original hundreds of millions of images it  used in training. there are many descriptions, of varying detail, on how this actually works - posted frequently here & available online, so there's really no excuse for continuing to promote this false idea.

the completely separate argument is whether there should be any payment for images to be used in training, but no one has been able to show that pieces of their image shows up in a new creation

To say it right away: I don't have any specific knowledge because I haven't experimented with AI yet. In this respect, I agree with you.

So I can only try to draw conclusions from what I can see.

Maybe I really don't understand the working principle of AI software.

But, if it were as you say, that AI does NOT use ANY elements from existing images, how can it be explained that, for example
- the Apple logo
- the Apple mouse
- the iMac
- the keyboard
- the Mercedes star
and so many other elements can be seen unchanged from the original in the images? Then why doesn't the AI "design" a new Apple or Mercedes logo, a different foot of the iMac, a new mouse etc.? I see here exactly the design features of Apple (material/color, radii, shapes, etc.).

And, if it is as seen here, who can rule out that elements from your and my pictures appear 1:1 in other pictures.

If you have a link that helps me to understand this, I would be grateful.

i was referring to knowledge about how ML works, not specific experience with using AI


this has been discussed many times - here's another example:
Interestingly, it obviously copies quite a bit as they were also including watermarks with the images they produce.
Might risk sounding like a broken record, but: The AIs sometimes generated images that have something resembling microstock agency watermarks, because they have been trained with so many watermarked (unlicensed!) images that they wrongly learned that the watermark was part of whatever it was supposed to generate. When an AI generates a watermark, it "thinks" it belongs in the picture like a suit to a businessman or the sun to a picture of a sunny sky. It's an issue of wrong learning, not an issue of copying. It recreates the watermark, just like it re-creates the sun or a suit. It cannot understand that the watermark is not part of whatever it is supposed to depict. If an AI was capable of thinking/realizing that whatever it is creating in images was actually something that exists in the offline world, then it would think that people walk around with floating watermarks in front of them.

I start to think that many people do not really understand what an AI is. Artificial intelligence. It's not a computer programm that copy & pastes stuff. It is a program that has learning abilities. It gets input and it learns from it. Give it the wrong input and it will learn to create wrong results.

thus trademarks appear not because these are copied from a particular image but because many images contain those TM, the TM becomes part of its knowledge of what a computer looks like. it extracts info & stores it in a different format so after training it doesn't know anything about the original images.


https://nanonets.com/blog/machine-learning-image-processing/#working-of-machine-learning-image-processing  gives a quick overview of what's involved. even at this high level it requires some mathematical knowledge which likely explains why so many posting here misunderstand what's happening, and why they wrongly think AI is taking parts of their images directly to create new images.

and to return to your question
And, if it is as seen here, who can rule out that elements from your and my pictures appear 1:1 in other pictures.

that's asking to prove a negative; instead one needs to show that the claim actually occurred. it's another example of a misunderstanding the actual process
« Last Edit: November 29, 2023, 12:35 by cascoly »

« Reply #57 on: November 29, 2023, 14:24 »
+3

thus trademarks appear not because these are copied from a particular image but because many images contain those TM, the TM becomes part of its knowledge of what a computer looks like. it extracts info & stores it in a different format so after training it doesn't know anything about the original images.


Steve, now we're getting more into the philosophical level  ;)
For example, if the AI regularly displays the Apple logo in a realistic form, then it is a copy. Apparently, AI is not capable of creating its own Apple logo.

If I now have a rare, unrivaled landscape image or an object in my portfolio, the AI has no choice but to copy it too  ::)

« Reply #58 on: November 30, 2023, 12:51 »
0
...
If I now have a rare, unrivaled landscape image or an object in my portfolio, the AI has no choice but to copy it too  ::)

sounds like you have a great experiment awaitin'

« Reply #59 on: December 08, 2023, 20:02 »
+1
Just moments ago i realize one thing...

It's curious no one in this forum (including me) talked about Metas new AI image generator was trained on 1.1 billion Instagram and Facebook photos. Maybe the focus somehow was just in our work than part in ourselves too but this got me thinking...

In the new world of AI: we are a customer or we are the product.

Certainly people can have different levels of AI use from zero to "all in" but definitely AI will use us. 
« Last Edit: December 08, 2023, 20:05 by Evaristo tenscadisto »


 

Sponsors

Mega Bundle of 5,900+ Professional Lightroom Presets

Microstock Poll Results

Sponsors