HOME | DD

Description
Finally, the day has come for our humble beginnings to unfold its final step of becoming a futuristic metropolis!
And to finalise our journey through time, I would like to talk about one of the most futuristic things that is going on in the world right now.

AI art.

This image was made with AI. Please read on if you would like to hear the side of an AI 'artist' (or, as I like to call myself, a 'prompt engineer').

I hear you. I think many people in the world hear you.
Just today, DeviantArt has posted new articles regarding AI art, and what you can do to protect your own artworks from any future data scraping.
You've worked so hard, learned so much, that it would of course be terrible for robots to come in and swoop your job away.

However, DeviantArt is working hard trying to protect you by enabling an (opt-out by default) setting for when you upload your images, which will indicate that Machine Learning models are not allowed to be trained on that image.
Moreover, I really don't think artists' jobs will go away, (I elaborate on this further on, but) essentially I think manual labour has value of its own.
Lastly, it enables more people to let their creativity roam free where they couldn't before; imagine not being able to physically draw, but with text-to-speech and AI text-to-image models, those people can now also create art!

But again, yes, I understand all the artists' fears. I think the most important thing is to adapt. Use it. See for yourself how it works, and how it could empower you.

It's not just a lifeless tool, a machine just churning out 100s of images all day long; it requires hard work to strike the right balance, find the right settings, create the right prompt, and eventually get the result that you like as well. It's not as easy as people make it sound to be, and there is still very much of a manual labour (and love!) that goes into making these pieces. I would say, for some more than others, but if I speak for myself there actually goes a lot of iteration into every single one of my pieces (imagine 100s of images being generated before I get to the result I want, and even then I usually have to combine it with GIMP, or even other latent diffusion models, in order to get a more polished result, e.g. removing unwanted extra fingers, or adding in new details like legs when there are none in the image that was generated).

But please, don't hate on AI.

AI art is also art.

There are real ideas and thoughts that went into creating these, just like a painter (or any other type of creator) would.

Is every image made in a fraction of the time? Yes. Does that scare many people? Of course! But I think that there will always be a market for hand-made creations, for example for process videos, or just because of the very fact that it is, indeed, hand-made. And, as I said above, is that it still takes a long time to get to a result that you like. Is one image made in a minute? Yes. Does the entire process of getting to an image that I would upload onto DeviantArt take one minute? Definitely not!
All I can say is that, for me as someone with a creative vision, but not the skillset to draw / paint (believe me, I've tried, for years, and I envy any and all of you who can make all these beautiful creations!). For many years I've tried to draw, paint, create digital 2D art, model 3D objects, renders, levels, games, etc. And now, with the advent of AI, another possibility opens up to empower my creativity.
What I would like to add onto that is that the world is fast changing. DeviantArt released an article just today about how you can now protect your own art by tagging it with a no-ai tag, so that machine learning systems can identify and, in theory, will not use any images which contain such tags in their metadata. Once things mature enough, and these systems are enforced by laws, it will hopefully be better. For now, if it is any comfort, you (the REAL artists out there!) are empowering people like me to create our visions. I cannot be more grateful for this wonderful technological feat that I daresay we, as a collective, have achieved. The sheer amount of data that has been used (more than 2 BILLION images in Stable Diffusion, and a newer Chinese model has even been trained on about 10 times that amount!) truly makes it a collective endeavour. I understand that for some (or even most) of the people whose data was taken, this is your lifelong work, you have studied and practised all your life, only to have an AI 'replicate' your work. However, one thing I would like to clarify, as many people seem to think that 'replicating' is exactly what it does. It does not copy any images. It simply couldn't. If it's a copy, that only shows the sheer technological marvel that has been achieved; the entire model learns what images look like based on their textual description, and creates weights out of all of this, effectively rewiring its 'brain', or better said, its neural pathways. To continue this anthropomorphic metaphor, AI is a painter who has seen a lot of pictures, and learned to identify many of their features in order to 'replicate' objects, styles, and many more; just like a normal painter would learn by observing a lot of pictures. I think that, without looking at any pictures, or any visual reference, an aspiring painter would have a hard time.
And, yes, the AI is really good at remembering all of those different features in all of those different images. Does it remember the images themselves? No. It knows them. It has learned from them. And now, it can make all kinds of things. Such as DnD maps! Characters! Environments!
I understand fully the 'threat' that AI poses. However, I think the key point is that it enhances us, and can truly empower artists. Please, try it out yourself, and see if you can find a workflow that enhances your work. For example, you can use it to inspire your creativity, or you can create a very rough first sketch with an AI, and edit that yourself with a program such as GIMP or Photoshop.
If anything, there are so many new possibilities, this is only the beginning of a wonderful new journey in human technology!

A New Directive for Opting Out of AI Datasets

Control The Amount of AI Art You See

DeviantArt has added the option to show less images that are made by AI. I have heard concerns of people who are asking, "but how will we counter the people that won't tag their art as AI art?": by simply making the AI software itself generate that metadata, rather than leaving it up to the end-user.
So, while currently, this may still be up to the decency of those who upload their images and how they tag them, in the future, images could automatically be tagged by AI that they are indeed AI images; therefore, easily filtered out by DeviantArt.

I can definitely see Stable Diffusion, DALL-E, Midjourney and more AI systems tagging their images with similar metadata in the future as well, so that these images can be filtered out more efficiently. This actually has a second reason: because they don't want to train new ML models on AI-generated data. By tagging it themself, they can also make sure that images that were indeed made by AI won't be used for training purposes.
And, moreover, to come back to the original question, this will make sure that AI images will not end up on your feed.

What's more, when the initial announcement above about opting out of AI datasets was released, you had to opt-out yourself with every image you uploaded. Now, DeviantArt has listened to the community, and made it an opt-out setting by default, meaning that all of your images should now automatically be 'protected' from machine learning models.

So again, I understand this is a scary time for all of you artists out there, but there is a lot going on in an attempt to protect your hard work!

More than that, I think now is the time to adapt. Go with the flow, see how you can use it. I'm sure all you creative people out there can come up with hundreds of creative ways to use this new technology! Don't simply discard it!

#AIArtIsAlsoArt

#AIArt

Note: This is not the original image. Would you like to see more of the process that goes into my pieces? Consider becoming a watcher!
Watchers will also get a timelapse video at the end of this journey through time!

Note about purchase: Includes both the original and the edited image.
SUBSCRIBERS GET FREE DOWNLOAD ACCESS IN A SUBSCRIBER-ONLY GALLERY.

Description
Finally, the day has come for our humble beginnings to unfold its final step of becoming a futuristic metropolis!
And to finalise our journey through time, I would like to talk about one of the most futuristic things that is going on in the world right now.

AI art.

This image was made with AI. Please read on if you would like to hear the side of an AI 'artist' (or, as I like to call myself, a 'prompt engineer').

I hear you. I think many people in the world hear you.
Just today, DeviantArt has posted new articles regarding AI art, and what you can do to protect your own artworks from any future data scraping.
You've worked so hard, learned so much, that it would of course be terrible for robots to come in and swoop your job away.

However, DeviantArt is working hard trying to protect you by enabling an (opt-out by default) setting for when you upload your images, which will indicate that Machine Learning models are not allowed to be trained on that image.
Moreover, I really don't think artists' jobs will go away, (I elaborate on this further on, but) essentially I think manual labour has value of its own.
Lastly, it enables more people to let their creativity roam free where they couldn't before; imagine not being able to physically draw, but with text-to-speech and AI text-to-image models, those people can now also create art!

But again, yes, I understand all the artists' fears. I think the most important thing is to adapt. Use it. See for yourself how it works, and how it could empower you.

It's not just a lifeless tool, a machine just churning out 100s of images all day long; it requires hard work to strike the right balance, find the right settings, create the right prompt, and eventually get the result that you like as well. It's not as easy as people make it sound to be, and there is still very much of a manual labour (and love!) that goes into making these pieces. I would say, for some more than others, but if I speak for myself there actually goes a lot of iteration into every single one of my pieces (imagine 100s of images being generated before I get to the result I want, and even then I usually have to combine it with GIMP, or even other latent diffusion models, in order to get a more polished result, e.g. removing unwanted extra fingers, or adding in new details like legs when there are none in the image that was generated).

But please, don't hate on AI.

AI art is also art.

There are real ideas and thoughts that went into creating these, just like a painter (or any other type of creator) would.

Is every image made in a fraction of the time? Yes. Does that scare many people? Of course! But I think that there will always be a market for hand-made creations, for example for process videos, or just because of the very fact that it is, indeed, hand-made. And, as I said above, is that it still takes a long time to get to a result that you like. Is one image made in a minute? Yes. Does the entire process of getting to an image that I would upload onto DeviantArt take one minute? Definitely not!
All I can say is that, for me as someone with a creative vision, but not the skillset to draw / paint (believe me, I've tried, for years, and I envy any and all of you who can make all these beautiful creations!). For many years I've tried to draw, paint, create digital 2D art, model 3D objects, renders, levels, games, etc. And now, with the advent of AI, another possibility opens up to empower my creativity.
What I would like to add onto that is that the world is fast changing. DeviantArt released an article just today about how you can now protect your own art by tagging it with a no-ai tag, so that machine learning systems can identify and, in theory, will not use any images which contain such tags in their metadata. Once things mature enough, and these systems are enforced by laws, it will hopefully be better. For now, if it is any comfort, you (the REAL artists out there!) are empowering people like me to create our visions. I cannot be more grateful for this wonderful technological feat that I daresay we, as a collective, have achieved. The sheer amount of data that has been used (more than 2 BILLION images in Stable Diffusion, and a newer Chinese model has even been trained on about 10 times that amount!) truly makes it a collective endeavour. I understand that for some (or even most) of the people whose data was taken, this is your lifelong work, you have studied and practised all your life, only to have an AI 'replicate' your work. However, one thing I would like to clarify, as many people seem to think that 'replicating' is exactly what it does. It does not copy any images. It simply couldn't. If it's a copy, that only shows the sheer technological marvel that has been achieved; the entire model learns what images look like based on their textual description, and creates weights out of all of this, effectively rewiring its 'brain', or better said, its neural pathways. To continue this anthropomorphic metaphor, AI is a painter who has seen a lot of pictures, and learned to identify many of their features in order to 'replicate' objects, styles, and many more; just like a normal painter would learn by observing a lot of pictures. I think that, without looking at any pictures, or any visual reference, an aspiring painter would have a hard time.
And, yes, the AI is really good at remembering all of those different features in all of those different images. Does it remember the images themselves? No. It knows them. It has learned from them. And now, it can make all kinds of things. Such as DnD maps! Characters! Environments!
I understand fully the 'threat' that AI poses. However, I think the key point is that it enhances us, and can truly empower artists. Please, try it out yourself, and see if you can find a workflow that enhances your work. For example, you can use it to inspire your creativity, or you can create a very rough first sketch with an AI, and edit that yourself with a program such as GIMP or Photoshop.
If anything, there are so many new possibilities, this is only the beginning of a wonderful new journey in human technology!

A New Directive for Opting Out of AI Datasets

Control The Amount of AI Art You See

DeviantArt has added the option to show less images that are made by AI. I have heard concerns of people who are asking, "but how will we counter the people that won't tag their art as AI art?": by simply making the AI software itself generate that metadata, rather than leaving it up to the end-user.
So, while currently, this may still be up to the decency of those who upload their images and how they tag them, in the future, images could automatically be tagged by AI that they are indeed AI images; therefore, easily filtered out by DeviantArt.

I can definitely see Stable Diffusion, DALL-E, Midjourney and more AI systems tagging their images with similar metadata in the future as well, so that these images can be filtered out more efficiently. This actually has a second reason: because they don't want to train new ML models on AI-generated data. By tagging it themself, they can also make sure that images that were indeed made by AI won't be used for training purposes.
And, moreover, to come back to the original question, this will make sure that AI images will not end up on your feed.

What's more, when the initial announcement above about opting out of AI datasets was released, you had to opt-out yourself with every image you uploaded. Now, DeviantArt has listened to the community, and made it an opt-out setting by default, meaning that all of your images should now automatically be 'protected' from machine learning models.

So again, I understand this is a scary time for all of you artists out there, but there is a lot going on in an attempt to protect your hard work!

More than that, I think now is the time to adapt. Go with the flow, see how you can use it. I'm sure all you creative people out there can come up with hundreds of creative ways to use this new technology! Don't simply discard it!

#AIArtIsAlsoArt

#AIArt

Note: This is not the original image. Would you like to see more of the process that goes into my pieces? Consider becoming a watcher!
Watchers will also get a timelapse video at the end of this journey through time!

Note about purchase: Includes both the original and the edited image.
SUBSCRIBERS GET FREE DOWNLOAD ACCESS IN A SUBSCRIBER-ONLY GALLERY.