Café de Nuit

The 'Seek permission'-part is nigh-constantly lacking. As far as I know, most of the AI scraping the net do not ask permission. Sadly, many hosting sites feature an opt-out policy for allowing techbros to 'train' AI on art, rather than an opt-in - and even then, there is a lot of abuse.

I recall horror stories of some techbro smugly telling an artist who'd honed their skills for years that they'd programmed their AI to copy his style, exclusively, as if he expected thanks or praise, then proceeded to start selling what his program vomited up.

Another artist got a message from people he'd done work for and billed appropriately. They'd seen what another techbro was charging, which was of course much cheaper. They demanded the artist "take a good long look at what's wrong with you" and scale down the bill.

And let us not forget the lovely "debate" between an artist and a genuine idiot over the definition of art. The artist said what the idiot was offering up was not his own work, because the computer had put it together for him based on general parameters. Quoth the idiot: "Well, your pencil made yours for you!"

With regards to 'successful artists not suffering from the technology so much' - I beg to differ. The better known you are, the more of your work is out there, the bigger the chance some joker will target you. And success requires being able to make money off your work. This becomes more difficult as people and companies keep going for the cheaper option.

I agree with the broken ethics of AI using previous stuff to create new one. I’m still not using Uber because I believe the tactics they used to impose themselves in the taxi industry were often bullying (at least here where I live).

But another thing you have to consider is its inevitability

Whatever option we take, AI imagery will remain and stay. So whatever we do, it will thrive.

As the biggest fan RL site, the FoS didn’t jump in the 5e RL bandwagon, and now in here we have few of those new 5e people that were drawn to Ravenloft, and that means we miss a lot of people.

Banning / hating 5e RL didn’t change a thing, it became popular with the 5e crowd.

Hating Uber didn’t change a thing, it became popular and imposed itself at the prejudice of taxi drivers.

About AI, I’m also sure we can’t change a iota in this: whatever our position, it will be used

So why not permit it, and make a code of use, such as (to be developed):

- We prefer real art made by real people. Really, if you can do this, you have bonus cookies and our appreciation.
- We accept good quality AI (to be developed), as long as it’s properly identified, and prompt is provided. Ex:

Made by AI using Chat GPT, prompt was “Generate the image of a pair of wolves in a forest trail, in front of the heroes.
They are feeding on the carcass of some black animal. The wolves snarl at the heroes.”

So that way we leave the choice to the writer.
You do not want to use it? Fine.
You want to use it? Fine.

Accountability would be nice.
'Make sure to only use imagery that is permitted for use'.

Rock of the Fraternity wrote: ↑Thu Jan 02, 2025 10:14 am Accountability would be nice.
'Make sure to only use imagery that is permitted for use'.

How do you do that with AI? If you ask this requirement, you de facto refuse AI images.

As Meph said, you can see it as a new tool, we can use it if we regulate its use. People said the same thing when Photoshop started.

In our quiet street, we prefer horse drawn carts, and if you use a car, reduce your speed

I have no money to pay for an artist. If I were to use art in a professional way I would pay for the artist, but for free products and friendly projectrs, then yes I agree.

Mephisto of the FoS wrote: ↑Thu Jan 02, 2025 1:15 am

IanFordam wrote: ↑Wed Jan 01, 2025 7:48 pm Then again, his wife then reminded him of another friend who had his copyrighted work recognizably regurgitated by an AI engine.

So is every collage in the world, and people make money on this and it is acceptable although legally illegal.

If you reproduce, publish or distribute a copyrighted work (or a work derived from a copyrighted work) without permission or a valid license – that's copyright infringement. If you want to use an image that's copyright protected, first get a license or permission to use it from the creator.

Often the materials will be copyrighted. So your unauthorized use of those materials would be copyright infringement unless your collage qualifies as fair use. Unfortunately, there is no legal rule on whether collage as a category would be fair use.

Any work published before 1924 is now in the public domain and can be freely used in your work. Any images created before 1924, are copyright free. So in other words any collage you see is probably illegal.

Free images are typically licensed with Creative Commons copyright licenses or are a part of the public domain. Public domain images and images with a Creative Commons Zero (CC0) license are free to download and use commercially without attributing the source.

Fair use permits a party to use a copyrighted work without the copyright owner's permission for purposes such as criticism, comment, news reporting, teaching, scholarship*, or research.

Accordingly, you cannot claim copyright to another's work, no matter how much you change it, unless you have the owner's consent.

Reposting someone else's copyrighted work on Social Media can open the door to costly copyright infringement lawsuits – even if you give credit to the copyright owner. Technically speaking, reposting another user's content — reposting social media content — is copyright infringement.

It's a rabbit hole...

*PS
We are scholars and researchers aren't we?

Isn't parody and satire included in the fair use?

I can't pay a human artist for work in a netbook distributed for free. And for all the "starving artist" talk, commissions of art can cost $1000. No guarantee that the artist will share your vision, even after the concept is fully explained as well. With AI I can generate hundreds of images and select the image that meets my needs. The AI models use data in aggregate rather than steal the works of any artist in particular (although, it is quite good at imitating styles when requested). For me, it's definitely a yes. I was thinking of creating TSR style trading cards using AI images.

Mistmaster wrote: ↑Thu Jan 02, 2025 8:27 pm I have no money to pay for an artist. If I were to use art in a professional way I would pay for the artist, but for free products and friendly projectrs, then yes I agree.

tomokaicho wrote: ↑Fri Jan 03, 2025 1:29 am I can't pay a human artist for work in a netbook distributed for free. And for all the "starving artist" talk, commissions of art can cost $1000. No guarantee that the artist will share your vision, even after the concept is fully explained as well. With AI I can generate hundreds of images and select the image that meets my needs. The AI models use data in aggregate rather than steal the works of any artist in particular (although, it is quite good at imitating styles when requested). For me, it's definitely a yes.

Well said.

tomokaicho wrote: ↑Fri Jan 03, 2025 1:29 am I was thinking of creating TSR style trading cards using AI images.

Nice.

As a writer, I can only extrapolate my feelings about AI imagery by comparing it to ChatGPT and other textbots. Basically, if I do the work I expect an attribution at the very least, if not necessarily any remuneration. With my QTR articles I have tried to be very scrupulous with the pictures I use (Creative Commons is a fantastic resource) and whose writings I'm using as source material. Sure, I'm not getting paid for my contributions, so there's really nothing at stake regarding copyright or royalties and such, but I want to give credit where credit is due. So, in the end, I'd say I'm on the “nay” side in regards to using AI art, even if I think my articles could benefit greatly from it.

Mistmaster wrote: ↑Thu Jan 02, 2025 8:27 pm I have no money to pay for an artist. If I were to use art in a professional way I would pay for the artist, but for free products and friendly projectrs, then yes I agree.

My take too.

Many of us are old enough to remember the days when Photoshop and other similar tools showed up 25-30 years ago and there was a wooha about the "real" artists being pushed aside too, back then. It is 2025 and digital tools for making art or touch it up, like photoshop and the like are not as "acceptable" as people drawing with pen and paper.
Still, there was talk about art dying in the late 90s because of digital tools and it didn't; it just changed. I expect the same will happen now.

tomokaicho wrote: ↑Fri Jan 03, 2025 1:29 am I was thinking of creating TSR style trading cards using AI images.

A very good idea. Especially for "friendly" games. I have a friend that likes, in D&D, to have his abilities and spells and all in cards. He looks at the net for the right image, crops it or reshapes and fits it on a paper to put the text in the back. It takes him hours but he claims it is very helpful. He seems to enjoy it so all the power to him.
Anyway, that would help him or people like him.

If I count correctly, so far, there is a good majority of Yay over Nay/not sure.

I understand that those on the Nay side are very much against it.

So not to force AI on anyone, why not give writers the choice? You refuse to use AI, fine. For whatever reason, you want to use AI, fine too. We do not force you to use it.

I propose again – as long as our stuff is distributed free :

If you want to add art to your article, you have the choice :

- We will always prefer real art made by real people.

- But if you prefer to use AI, we also accept good quality AI*, as long as it’s properly identified, and prompt is provided. Ex:

Made by AI using Chat GPT, prompt was “Generate the image of a pair of wolves in a forest trail, in front of the heroes. They are feeding on the carcass of some black animal. The wolves snarl at the heroes.”

* that is, we will refuse AI with obvious flaws such as weird body size or round / odd face, missing nose or limb, weird number of fingers (or melted or too long), etc. Do not submit these. Try again to remove these flaws or use Photoshop to correct it. If you pastiche / parody a well known piece of art, identify the original ex. based on "XYZ" by ABC. Do not submit an image that is based on a famous living person (ex. George Clooney as a pirate).

What do you guys think?

I agree, if my opinion is requested.

I feel that it's slop that degrades all it touches. It feels so superficial and artificial. I put up with it in Threshold Magazine, but I glaze over AI generated pieces much of the time.

I'm definitely more on the side of the "nays", and I echo the sentiments brought up by Rock, Ian Fordam, Jeremy16, and more. So in order for me to really contribute anything of value to the thread, I want to highlight resources that aren't AI but are still of great use if one doesn't want to commission an artist. Jeremy16 brought up Creative Commons so I won't do that.


If you're ever looking for inspiration from real life paintings, I really recommend the following: https://www.artnet.com/net/galleries/se ... works.aspx & https://www.nga.gov/collection/collection-search.html & http://www.artcyclopedia.com/ & https://www.wikiart.org/. They're relatively easy to navigate, give or take a few tries and errors, and with them you can find excellent portraits and art pieces to evoke the exact feeling of Ravenloft that you want in your games. I used these bad boys for Ghastria, to show off the art within certain art galleries and to use certain paintings of everyday life from impressionist artists as set-pieces for when I didn't need to break out the battlemap and ensuing grid! They can do similarly for any netbook as well.

Additionally, I know that DeviantArt was brought up already, but I have to also say some very positive things about ArtStation (https://www.artstation.com/). If there's any video game or concept art from a film or book or whatever that you thought fits Ravenloft, you can use ArtStation to find those and plug them into your games. In my case, I go find concept art or level design from developers on Art Station of games that I think fit Ravenloft. I need pics for Pharazia? Good thing all the concept art, game art, level design, environmental assets and more for Assassin's Creed: Mirage came out not too long ago, so I can just take those JPEGs and port them over to Roll20 and voila! Or put them in my "Campaign Relevant Images" channel on my Ravenloft player/party Discord server, easy peasy. Other sites that do very similarly to ArtStation would be CreativeUncut (https://www.creativeuncut.com/) FZDSchool (https://fzdschool.com/galleries), and probably more.

I will also say that there's a lovely appeal to finding the right images that people believe is "Ravenloft" at its core, and being able to show them off. A significant chunk of the FoS user base loves Talon Dunning's work, and the covers of 2e adventures also did a great job of portraying the mood.

But speaking as someone who came in with 5e, and heard and either agreed or disagreed with the many criticisms of the 5e book that, art was definitely brought up. And as I see more and more people get into Ravenloft through the 5e book (I'm in the server for it), they have their own art channel. And more importantly, they all have their own unique conceptions of Ravenloft. Everyone does/did, whether they started with VRGTR or the Black Box. With that, I can fully get behind there being less art to begin with, if only because I wasn't relying on a wealth of art to convey Ravenloft to me the same as it would be conveyed to, say, you. I started with 5e and enjoy the 5e artwork. Not all of it, but for Ravenloft's case, there are some images that I do very much enjoy
But I also look back at 2e/3e art and I find there to be something else in there that I also enjoy.

Joël of the FoS wrote: ↑Sat Jan 04, 2025 11:33 am * that is, we will refuse AI with obvious flaws such as weird body size or round / odd face, missing nose or limb, weird number of fingers (or melted or too long), etc. Do not submit these. Try again to remove these flaws or use Photoshop to correct it. If you pastiche / parody a well known piece of art, identify the original ex. based on "XYZ" by ABC. Do not submit an image that is based on a famous living person (ex. George Clooney as a pirate).

What do you guys think?

I like what you put out Joël in regards to the prompt being put in as well, and I'm mostly in agreement. But there's another thing I'd really like to add, and it's about the glossy sheen of AI artwork. The lighting is so off that it draws me out of the image and it's so clearly AI because of how glossy it is. If there's any way to fix that, it would be really appreciated. Same with stuff like the clothes blending into themselves and buildings in background images too.

Lucius wrote: ↑Tue Dec 31, 2024 11:33 am

Joël of the FoS wrote: ↑Tue Dec 31, 2024 10:37 am * But if there are people reading this that wants to contribute art, please manifest yourselves

I can contribute one thing or another. Just let me know what you need.

Check your PMs