AI generated art

[Astro Pen]

Irony abounds. I looked at this article and Ralph the doggy is sufficiently cute, in the first picture, that my first thought was a CGI . But no, he is real.

Wrexham: Missing dog takes 100-mile round-trip in taxi

A dog jumps into a cab after running off on its morning walk and takes a trip to Manchester airport.

www.bbc.co.uk

 

[Stephen Palmer]

Agreeing with the above - images, not art.

AI Art & All That

I’ve been very interested in the current debate about AI art, which has kicked off in recent months. I’ve read lots of thought-provoking blogs, Facebook posts and more, which have pushe…

wp.me

 

[Mon0Zer0]

From @Stephen Palmer 's article:

In my view then, the role of the artist in using AI software such as Midjourney is choice. Midjourney offers iterations – a number of image variations. A sufficiently experienced and talented artist can choose the “best” ones. That is an artistic act in my view. It could be argued that anyone can make such a choice, and that would be correct, but the artistic eye not only has to be rooted in true creativity, which not everyone has in useable amounts, but it also has to be trained. There is therefore a spectrum of curation available, depending on the artist’s gift.

Spot on - although I would add that the skill now is in language and artistic knowledge to be able to create a prompt. Prompt engineer is now a recognised and lucrative job - crafting images by using words and drawing on art history.

 

[Stephen Palmer]

Spot on - although I would add that the skill now is in language and artistic knowledge to be able to create a prompt. Prompt engineer is now a recognised and lucrative job - crafting images by using words and drawing on art history.

What an irony!

 

[Wayne Mack]

From @Stephen Palmer 's article:



Spot on - although I would add that the skill now is in language and artistic knowledge to be able to create a prompt. Prompt engineer is now a recognised and lucrative job - crafting images by using words and drawing on art history.

The reality is that current AI text deteriorates rather quickly following the initial prompt. The AI does not have an overall arc derived from the prompt, rather it is continually guessing at what is next. Also consider that even the most basic short story will have something more than a single plot line. In addition to whatever is directly related to the initial prompt, there are also subplots, character arcs, and world building aspects.

Even discovery writers will have some basic overall feel or tenor that they will try to create. This provides some boundaries and constraints on what will happen in their stories. AI technology doesn't have this and will simply wander following the prompt and will not necessarily stay within the boundaries of the initial prompt.

To paraphrase Mark Twain, it is not so much that AIs are doing well at creating text; it is just interesting that they can do it at all. The combinatorial issues when using further look ahead or future planning in AI tech quickly outstrips any known or projected technologies to support it. I think the hype cycle around AI has peaked (again) and the realization is setting in that it has a lot of limitations and can (and does) result in some major consequences. There are no AI-driven novel factories in our future.

 

[Mon0Zer0]

The reality is that current AI text deteriorates rather quickly following the initial prompt. The AI does not have an overall arc derived from the prompt, rather it is continually guessing at what is next. Also consider that even the most basic short story will have something more than a single plot line. In addition to whatever is directly related to the initial prompt, there are also subplots, character arcs, and world building aspects.

We're talking about AI Visual Art, not Chat GPT.

 

[Robert Zwilling]

There is an old saying, one picture is worth a thousand words.

 

[Mon0Zer0]

The reality is that current AI text deteriorates rather quickly following the initial prompt. The AI does not have an overall arc derived from the prompt, rather it is continually guessing at what is next. Also consider that even the most basic short story will have something more than a single plot line. In addition to whatever is directly related to the initial prompt, there are also subplots, character arcs, and world building aspects.

Even discovery writers will have some basic overall feel or tenor that they will try to create. This provides some boundaries and constraints on what will happen in their stories. AI technology doesn't have this and will simply wander following the prompt and will not necessarily stay within the boundaries of the initial prompt.

To paraphrase Mark Twain, it is not so much that AIs are doing well at creating text; it is just interesting that they can do it at all. The combinatorial issues when using further look ahead or future planning in AI tech quickly outstrips any known or projected technologies to support it. I think the hype cycle around AI has peaked (again) and the realization is setting in that it has a lot of limitations and can (and does) result in some major consequences. There are no AI-driven novel factories in our future.

One of Game of Thrones producers had this to say the other day in reaction to the forthcoming WGA strike, re: AI and writing:

"The biggest threat is not the studios. It’s Ai. The apps - and where they are on their evolutionary track ALREADY pose a threat to all writers in the (quite near) future, and the studios know it and know how to profit from it. Soon, they won’t need 90% of writers. These Ai writing apps - chatgbt, sudowrite, etc., are still in relative infancy. Moving fast toward the level of structured, functional creative product which most shows feature these years.


The best stuff, sure. There’s still time for the best of writing in any category. But for the majority? It’s a tsunami and is revolutionizing every aspect of our industry NOW.

Be careful, strikers. You may be looking the wrong direction."

In response, an IP scout / Agent replied:

"I spoke to a German producer yesterday who said that they are embracing it...he had recently attended a lecture entitled "How to Use AI as a Staff Writer."
sh*t is about to get weird."

 

[Daysman]

Any one tried the bing app powered by dall-e?

Prompt: photorealistc spacecraft , pkd sci fi paperback cover art, style of chris foss

Don't know if it picked up on pkd... or Foss really. Colourful. . .

 

[Vertigo]

Here's a question I don't think I've seen asked but apologise if I have. If you use one of these 'AI' powered art generators to produce an image, who owns copyright of that image? So, thinking of @Daysman's example above, if I use an 'AI' to produce an image that I then use as a book cover on a commercially sold book am I legal?

 

[Wayne Mack]

Here's a question I don't think I've seen asked but apologise if I have. If you use one of these 'AI' powered art generators to produce an image, who owns copyright of that image? So, thinking of @Daysman's example above, if I use an 'AI' to produce an image that I then use as a book cover on a commercially sold book am I legal?

Maybe there is a use for NFTs!

 

[Robert Zwilling]

who owns copyright of that image

It appears that the US and Europe have 2 different ways of handling this, and apparently totally opposite.

The way things stand now, in the US, the program that made the image can not copyright the image. The image itself apparently can also not be copyrighted because it was not made by a human.
U.S. Copyright Office Says AI-Generated Images Do Not Qualify For Copyright Protection

Since it is not your program, you haven't bought it or created it, there are probably terms of service which have somehow become legal instruments that can govern how the users of the software can use the results. It may not be legal but a court would have to answer that question.

There are legal cases waiting to be resolved where the owners of the original art that was sampled (and originally copied to obtain the use of) are suing makers of the AI generated image software because the original owners never gave their permission for the artwork to be used.
Getty Images is suing the creators of AI art tool Stable Diffusion for scraping its content.

That probably stems from the terms of service agreements that handled the posting of the original images. The digital generation's (no age limits) answer to this is that if it ain't nailed down, it's free to take. (see fakebook for how this is used to make money)

I don't expect it to stay like this for 2 reasons.
1) The corporations that will be creating and using AI generators to create corporate work will not be happy if they can't control the process or own the results.
2) The US courts no longer make decisions based on modern precedent, or even existing precedent, which means anything goes.

Most people don't seem to understand that copying an image posted on the internet without permission from the owner is copying an image without permission from the owner.

This apparently is Europe's take on the issue:
‘In the case of a literary, dramatic, musical or artistic work which is computer-generated, the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken.’ — Section 9(3) CDPA 1988.

The article then goes on to say:
As far as AI is concerned, the question of authorship/ownership of AI generated work will hopefully be resolved in the next few years in the UK, which last year (2021?) published its National AI Strategy.


"In the meantime, disputes arising from computer generated works in the UK from software/‘AI’ — such as Stable Diffusion or DALL-E — will need to be assessed on a case-by-case basis and will likely depend on the level of contribution of each of the parties in the creation of the work."

They do have some advice which seems to go under the heading of don't ask, don't tell,
"Consider not using any artist’s name or work in the generation of artwork — especially where their work is still under copyright protection. Without knowing the specifics of the tools and prompts used, it’s difficult to give a definitive answer on whether using the style of an artist would be considered an infringement of their rights — but it may nevertheless annoy said artist and result in unwanted bad press and/or infringement proceedings."

If the book cover is for a digital only book, it will be very simple to change the image if things don't go your way. If it is a hard cover book being printed in advance, changing the image could be a problem.

I would assume that some people are creating AI generated images and simply saying that they made the image. Problem solved.

 

[Vertigo]

It appears that the US and Europe have 2 different ways of handling this, and apparently totally opposite.

The way things stand now, in the US, the program that made the image can not copyright the image. The image itself apparently can also not be copyrighted because it was not made by a human.
U.S. Copyright Office Says AI-Generated Images Do Not Qualify For Copyright Protection

Since it is not your program, you haven't bought it or created it, there are probably terms of service which have somehow become legal instruments that can govern how the users of the software can use the results. It may not be legal but a court would have to answer that question.

There are legal cases waiting to be resolved where the owners of the original art that was sampled (and originally copied to obtain the use of) are suing makers of the AI generated image software because the original owners never gave their permission for the artwork to be used.
Getty Images is suing the creators of AI art tool Stable Diffusion for scraping its content.

That probably stems from the terms of service agreements that handled the posting of the original images. The digital generation's (no age limits) answer to this is that if it ain't nailed down, it's free to take. (see fakebook for how this is used to make money)

I don't expect it to stay like this for 2 reasons.
1) The corporations that will be creating and using AI generators to create corporate work will not be happy if they can't control the process or own the results.
2) The US courts no longer make decisions based on modern precedent, or even existing precedent, which means anything goes.

Most people don't seem to understand that copying an image posted on the internet without permission from the owner is copying an image without permission from the owner.

This apparently is Europe's take on the issue:
‘In the case of a literary, dramatic, musical or artistic work which is computer-generated, the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken.’ — Section 9(3) CDPA 1988.

The article then goes on to say:
As far as AI is concerned, the question of authorship/ownership of AI generated work will hopefully be resolved in the next few years in the UK, which last year (2021?) published its National AI Strategy.


"In the meantime, disputes arising from computer generated works in the UK from software/‘AI’ — such as Stable Diffusion or DALL-E — will need to be assessed on a case-by-case basis and will likely depend on the level of contribution of each of the parties in the creation of the work."

They do have some advice which seems to go under the heading of don't ask, don't tell,
"Consider not using any artist’s name or work in the generation of artwork — especially where their work is still under copyright protection. Without knowing the specifics of the tools and prompts used, it’s difficult to give a definitive answer on whether using the style of an artist would be considered an infringement of their rights — but it may nevertheless annoy said artist and result in unwanted bad press and/or infringement proceedings."

If the book cover is for a digital only book, it will be very simple to change the image if things don't go your way. If it is a hard cover book being printed in advance, changing the image could be a problem.

I would assume that some people are creating AI generated images and simply saying that they made the image. Problem solved.

Interesting; not unsurprisingly it seems to be a bit of a grey area right now.

 

[Daysman]

Had another stab having found a more complex prompt in an article. Didn't get their result so modded it, mostly adding details to correct errors.

Prompt: A young beautiful girl wearing a necklace, earrings and a hat, pastel, rounded lips, shaped eyebrows, interesting eyes, slightly smiling. Modern. Photorealistic image, sharp focus, cinematic. Photograph using 50mm lens at f/4 and ISO 100:

Still can't get a good starship out if it though...

 

[Daysman]

Heh, what d'you know, starships . . . kinda . . .

 

[Toby Frost]

I'm not sure how much AI and how much human input is involved in this, but it did amuse me:

 

[AllanR]

I fed chat-gpt a bunch of metaphysics and asked it for a story proposal. I had to refine my prompts and edit after.

The Eternal Traveller
Summary: In a universe that is infinite, isomorphic, and has zero curvature, all the possible places and times coexist together. From the perspective of a photon, there is no travel time, and a being in the form of light could instantly travel to all the possible places and times. The protagonist, a traveller who can take the form of light, embarks on a journey to explore the concept of eternal recurrence, as theorized by ancient cultures and philosophers.
As the traveller navigates through different eras and civilizations, they encounter versions of themselves and other individuals from their possible and actual pasts and futures. They witness the rise and fall of civilizations, the birth and death of stars, and the evolution of life on different planets. They learn that every possible combination of forms must repeat itself eternally across both time and space.
Through their journey, the time traveller begins to feel trapped in this eternal cycle, and they seek a way to break free from it. They encounter a group of philosophers who have discovered how to manipulate the curvature of the universe, which allows them to create a portal to another dimension.
The traveller enters the portal, hoping to escape the cycle of recurrence, but they find themselves in a world that is completely different from anything they have ever known. As they explore this new world, they realize that they have the power to shape its future and create a new destiny for themselves.
In conclusion, "The Eternal Traveller" is a story that explores the concept of eternal recurrence and the possibility of breaking free from it. It takes the reader on a journey through time and space, and challenges them to question the nature of reality and their place in the universe.

 

[Foxbat]

I was pottering about on the internet kind of half-heartedly looking for any news on the next Dune movie when I came across this. I'm kind of wary of AI and the effect it will have both on artists in general and life as we know it.

That all being said, I love these images. Whether Fritz lang would have created anything similar is debateable (although I do think there is more than a hint of Metropolis here) but I would love to see a Dune movie with these kinds of set and costume designs - kind of Art-Deco Dune

 

[AllanR]

Professors, by field, according to Midjourney.

 

[KGeo777]

I finally tried an AI image generator.
Jeepers. It's cool and scary at the same time.