D&D General DALL·E 3 does amazing D&D art

[Jfdlsjfd]

The larger issue the AI doesn't have infinite images/ material to work from. It uses what is out there. Unfortunately, a lot of that is from copyrighted material that companies / people post online (often without indicating it is copyrighted material).

I've never, not once, generated anything that is even remotely like what is online and copyrighted. Largely this is because I use very long, specific prompts, which avoid AI just manipulating something that is copyrighted.

You don't need to do that. The AI doesn't interpret your prompt to select an image in a database and add or remove elements. The model doesn't contain any image. Even a small prompt will give you something different from what it was trained on, but there is always a chance, like the one million monkeys typing a piece from Shakespeare, that it will generate something identical to an existing artwork. The chance increase if you're prompting specifically to skew the generation toward it. But it will need a lot of conscious effort to infringe copyright, hence the copyright infringement being on the user.

Also, you could statistically recreate a copyrighted image even if the AI engine was trained only on public domain images or with image it was licensed to, like Adobe's. Prompt with a close enough depiction of Mickey Mouse without mentionning its name and you could certainly acheive something that is close enough to be infringing copyright after enough tries.

While the article was certainly a good one, many of the prompts are SO generic that the AI uses what it has the most samples of, which is typically copyrighted material.

Try it. Generate 1,000 images with "a painted portrait of a man". It's difficult to imagine something more generic. Chances of you replicating a real portrait of a man it was trained on are close to nil. Even if you try to skew it by saying "a 17th century portrait of a man" it will certainly not generate something identical to an already existing 17th century portrait it was trained on.

On the other hand, a very short prompt "Jack Sparrow" might very well generate something recognizable as Jack Sparrow. But if the user use that as a prompt and redistribute the work, I'd say it's his fault...

So, it doesn't surprise me the onus will fall on the user. The company can certainly do what it can, but it offers a service and what anyone chooses to do with that service in really on them, particularly on anything they then use commercially.

Much like the paint vendor doesn't get sued when you use their paint to paint Mickey Mouse and sell it. And you don't get any warning from Microsoft that using word implies the risk of infringing copyright (should you try to write fanfiction about an horrible soapy teenager love story between a werewolf and a vampire).

 

[Jfdlsjfd]

That is unreasonable, given than the AI can reproduce copyrighted material even when the user didn't ask it to. The user cannot reasonably be aware of every piece of copyrighted content in existence so that they could recognise when the AI has done so.

Note that generating a copyrighted image for your private use wouldn't be a problem. Copyright infringement starts when you're distributing the copyrighted material. There is nothing that preclude you from drawing a Mickey cartoon for your own private use. If you try to sell it, then as a professional it's reasonable IMHO to ask you to do the necessary diligences.

 

[ezo]

You don't need to do that. The AI doesn't interpret your prompt to select an image in a database and add or remove elements. The model doesn't contain any image. Even a small prompt will give you something different from what it was trained on, but there is always a chance, like the one million monkeys typing a piece from Shakespeare, that it will generate something identical to an existing artwork. The chance increase if you're prompting specifically to skew the generation toward it. But it will need a lot of conscious effort to infringe copyright, hence the copyright infringement being on the user.

Also, you could statistically recreate a copyrighted image even if the AI engine was trained only on public domain images or with image it was licensed to, like Adobe's. Prompt with a close enough depiction of Mickey Mouse without mentionning its name and you could certainly acheive something that is close enough to be infringing copyright after enough tries.

I know that. Perhaps the term "manipulated" I used was mis-leading. Sorry about that.

Exactly, I agree it should be on the user. People won't like it, but such is life.

Try it. Generate 1,000 images with "a painted portrait of a man". It's difficult to imagine something more generic. Chances of you replicating a real portrait of a man it was trained on are close to nil. Even if you try to skew it by saying "a 17th century portrait of a man" it will certainly not generate something identical to an already existing 17th century portrait it was trained on.

On the other hand, a very short prompt "Jack Sparrow" might very well generate something recognizable as Jack Sparrow. But if the user use that as a prompt and redistribute the work, I'd say it's his fault...

Those are still what I consider short prompts, so I would expect them to produce something the might be copyrighted. My prompts are typically about 500 words or more.

And yes, if the user purposely uses a short(ish) prompt and redistributes something likely copyrighted, that is totally on them.

Much like the paint vendor doesn't get sued when you use their paint to paint Mickey Mouse and sell it. And you don't get any warning from Microsoft that using word implies the risk of infringing copyright (should you try to write fanfiction about an horrible soapy teenager love story between a werewolf and a vampire).

Right.

 

[FitzTheRuke]

I have a very simple solution: Don't use AI for material that you plan to SELL. Employ Humans for that purpose.

Human artists should be able to use AI art for references and layout ideas, but not for finished art that they plan to sell. They should complete that art using their own skills. This gatekeeps for actual talent, which I think is a fair thing to do.

I also think that these AI art programs are pretty decent tools to teach human artists. I've been thinking of using it myself to practice a few cartooning angles that I've long struggled with.

 

[Saracenus]

I will just say that AI and our IP laws in the USA are going to have a reckoning in short order (and by short order via the courts I mean the next 5 to 10 years) because it is being added to everything.

• Adobe Suite has generative AI.
• Grammarly is using AI.
• Doctor case notes transcription is moving towards AI.
• AI art is creeping into the Stock Image collections and is hard to detect that it is hard to say that it is AI free, you just don't know.

The list keeps growing larger everyday. The idea that you are going to be able to avoid it is getting slimmer and slimmer.

Right now, I am not going to use any AI tools that would involve a commercial purposes because there is just too much risk (from being sued for infringement to not being able to copyright AI generated materials). Until AI and IP gets settled, this is a hobby and for personal use for me.

This is before you even get into the ethics of how these Large Language Model AI were created and trained in the first place. Or, the prospect of AGI (Artificial General Intelligence) which may or may not have already happened in the lab.

For an examination of the ethics and AI, the YouTube channel Philosophy Tube did a whole episode on it:

Spoiler: Philosophy Tube: Here's What Ethical AI Really Means

Meanwhile I will continue to create material for my own personal use and show off the good, the bad, and the ugly to you all.

 

[ezo]

I have a very simple solution: Don't use AI for material that you plan to SELL. Employ Humans for that purpose.

I agree... if you have the money to afford to pay humans to develop your artwork, etc.

I know, personally, I don't. So, anything I make for commercial use (at present) will probably use AI artwork. Now, of course, AI is far from perfect. For me, in that sense, it is a stop-gap. I use it now because I can't afford to pay digital artists for all the artwork I need.

Once (if LOL!) I ever make enough money, then in the future I would use the AI for conceptual (as you said), but hire people for the finished product so I can (hopefully) get exactly what my vision is.

 

[FitzTheRuke]

I agree... if you have the money to afford to pay humans to develop your artwork, etc.

I know, personally, I don't. So, anything I make for commercial use (at present) will probably use AI artwork. Now, of course, AI is far from perfect. For me, in that sense, it is a stop-gap. I use it now because I can't afford to pay digital artists for all the artwork I need.

Once (if LOL!) I ever make enough money, then in the future I would use the AI for conceptual (as you said), but hire people for the finished product so I can (hopefully) get exactly what my vision is.

Though I'm not sure I like it, I'm certainly far less concerned with those struggling to get going than I am with higher-end commercial use.

None of it would be a problem if most people didn't have to work really hard to get ahead while a select few enjoy the fruits of their labors, but that probably starts to get into territory that we're not allowed to discuss here.

 

[FitzTheRuke]

So! I've been working on making some "groups" of "monsters" (though usually more like NPCs) so that I can do larger-group combats without bogging my D&D game down. Without going into too many details for our purposes here, it's essentially using the "swarm" monster rules, but with people. Here's some of my icons, made using DALLE3:

(They've come out a little blurry when I transfer them here from the program I'm using, sorry about that)

 

[Andvari]

 

[Jfdlsjfd]

Though I'm not sure I like it, I'm certainly far less concerned with those struggling to get going than I am with higher-end commercial use.

I am pretty sure the higher-end commercial use will not be troubled with IP. They'll train on models that they got the right to by licensing (or outright buying) a stock photo website, and for Hollywood I a pretty sure there will be quickly enough retiring actor that will be glad to sell a perpetual right to their image in exchange for a luxurious end of life. Or talentuous hobbyists with a main job that would gladly sign a deal for rehearsal training of an AI every week-end for a few month in exchange of a big amount of cash... Once you have an AI able to recreate you, and to generate consistant shots of an edited real actor, they will need only the rights to 10 to 20 actors to do full CGI AI-acting for the whole industry. The tech might not be (fully?) ready yet, but it will be refined by the time the cast is assembled and the deal the actors got that prevented using image without selling it before will become irrelevant. The lawmakers will either strengthens copyright holder's right (and AI generation will generate profits mostly for the big names) or make training available to all and make everyone including the little ones able to profit from it. The result will be irrelevant for the jobs currently held and facing the prospect of being replaced, because it will be the same either way (in many if not most fields, including intellectual jobs and especially intellectual jobs).

None of it would be a problem if most people didn't have to work really hard to get ahead while a select few enjoy the fruits of their labors, but that probably starts to get into territory that we're not allowed to discuss here.

Indeed. And let's not blame the AI for problems that are coming from the taboo topic.