NEVER Plagiarize - EASILY use AI to create original artwork based off YOUR ART for you to reference
HTML-код
- Опубликовано: 3 окт 2024
- How to use AI to create original artwork from references you created so you can reference it. Confused yet? Watch the video 🤣 it will make sense soon.
Are you a digital artist who struggles with using references without feeling like you are copying someone else's work? Do you want to learn a cool and easy way to generate your own references using AI? If so, this video is for you!
In this video, I will show you how to use an app called Krea.ai, which allows you to transform any screen into an image using natural language prompts. You can use this app to create references from your own sketches, photos, or other sources, and then use them as guides for your artwork.
You will learn how to:
Set up Krea AI and select your preferred painting program
Use different prompts to generate different styles and effects
Adjust the size and position of the generated image
Save and export your references and artwork
By the end of this video, you will be able to create original and unique artwork using AI as a tool, not a crutch. You will also have more confidence and creativity in your artistic process.
If you enjoyed this video, please give it a thumbs up, subscribe to my channel, and leave a comment below. Thank you for watching and happy painting!
krea.ai
Join this channel to get access to perks:
/ @tabletpro
⭐ - Install the Tablet Pro STUDIO PEN & TOUCH apps : on-screen keyboard shortcut buttons and stylus customization - ⭐
www.tabletpro....
Keyboard shortcuts displayed and text expansion done with KeyPress OSD
www.keypressos...
Top Products List:
Drawing - REBELLE 7 PRO - www.tabletpro....
Stylus - Renaisser R530 - www.tabletpro....
Keyboard - Surface Pro 8 / 9 Bluetooth - amzn.to/3cOJjaf
Tablet :
Surface Pro 9 - amzn.to/3EgqbvU
Surface Laptop Studio 2 - amzn.to/3teYthn
Rog Flow Z13 - amzn.to/47Nfgrf
learn how to use Tablet Pro in the Tablet Pro academy
www.tabletpro....
It's literally the same korean model face from all the AI pics lol
This is the most honest way to use AI but, being you the one who's talking, there is no doubt about your choices. The only thing I'm concerned about is the fact that AI programs are fed and trained with images programmers rip from the web and pay not for.
It is interesting you use the word plagiarism in this context. You are stepping into a complex moral minefield here. As it stands right now most image generators are trained on all visual reference on the whole internet, which of course includes copyright protected content. In fact Midjourney have been exposed as specifically naming current artists work and offering them as styles. Others use the same datasets.
Now we know artists have been using other peoples work as reference forever, and that appeared to be acceptable artist to artist but now that it has reached industrial strength and fully automated it has caused a moral dilemma. If I can run a prompt that gives me a drawing in a live artists style is that okay even as in your example I then trace it and modify it. Not using an artists name in the prompt does not get away from the fact it does reference living artists work. (Albeit regenerated)
I think the law allows this traditionally so long as it is discernibly different to the original. Do we apply this to generative art? I guess it comes down to the individual, am I trying to make an original piece where the generated art is the starting place, or am I copying someone else's style because I'm too lazy to come up with on my own?
Maybe one day we will have a choice to use generative art that is entirely based on copyright free work. After all that is the whole of mankind's endeavours prior to 1954. (In the UK) Adobe and others are starting to say that they do but it isn't clear to me yet. Is Krea? I doubt it. I'm not judging anybody here, I'm still undecided on my own direction and as such am interested in the debates.
I agree with you. It is complex to say the least. I have a lot of questions as well. If the rip is a 90% - 95% similarity with a specific artists style, I think we start really getting into the grey. As I think most of our artwork really stays about 80% similar in all aspects, that remaining 20% or so becomes the relevant part. (my guestimates on percentages to illustrate a thought process) If two great artists create artwork now that is 98% similar but we as trained viewers of those can see which artist is which... well,... copyright becomes a real issue to uphold, especially stylistic copyright not individual works. So if a machine learns by aggregate, let's say from 1000 artists and from 1,000,000 photos (I would assume 1000 times that in reality for a good result) that the sum total of stolen material from a single artists "style" is going to be so small that it becomes indefensible in court.
Now,,... all that said. If they are specifically targeting various artists, or if a band of artists made their own ai trained model or could sell "models" based off their own work plus generalized ai structure... this could be very interesting as a market. I would like to see artists getting paid for their work. unless a specific prompt is allowed to almost perfectly replicate an artistic style I think it will be very hard to make any case.
Meaning, if we want to win a battle on this front, I think artists need to come together and make their own market place inside an ai platform.
this is still using copyrighted material
I know what you are saying but it's factually untrue. The models do use copyrighted material in their training but the generated images themselves are NOT copyright protected and in the USA not copyrightable by law. This is a moral maze for digital artists. Adobe and others are attempting to make their models free from copyright material and we all long for that, but Adobe will want top dollar for that. My hope is an open source group create a trained model based soley on copyright free content. In the UK that is anything where the originator has been dead for 70 years. I think in the USA that is 100 years.
@@uk3dcom its not untrue, like you said they are scrapping the net and using copyrighted material and material used without authorization to teach their AI, then they monetize the results. If this was free I would not complain but if they do business with it then they need to pay the artist they scrapped. Those AI scrappers are greedy corporates.
@colorfulbleeding I think we are arguing in agreement. It's probably the semantics.
We agree that the data sets have used copyrighted images in their training and that becomes a problem on many levels. What I was saying was that Justins use of a generated image wasn't infringing copyright. That in itself is a weird moral dilema, the generated image isn't by law copyright protected, but it has been produced through training that refrences copyright material....
How do we pick the bones out of that?
Hi Justice. Very informative as always, but I noticed that the sound keeps dropping out. That has not happened on your other videos.
there is a problem on my system where system sounds are interrupting the microphone input. Not sure the fix yet, but I did turn off the notification sounds this morning so hopefully that fixes it.
This is still not very useful unless you want something generic that you can find in a google search. That’s all it generates. If you want a different pose that is not the most common photo or drawing it can’t do it.
are you using your own sketch as the starting reference?
@@TabletPro Yes, don’t get me wrong, this is a fun software and it’s cool that you are highlighting it. I tried with a sketch of a tiger climbing up. The results this generates are your typical panther walking towards you that I have seen many times before, it goes crazy and but legs and tails everywhere.
@@TabletPro they're saying it's a glorified google search, which is accurate
What happens if you turn the ai strength down. For me it’s almost exactly what I’m drawing but modified
@@TabletProIf I turn it all the way down it looks like deformed version of what I drew. Again, not very useful. It’s not like it can give me a controlled style.
Good artwork btw!
I have a realistic style, so I reference photos, not other artists. But I still find it hard to find references with the right posing and angles.
yeah, that's why I like this so much. draw a very very loose sketch and then let the ai create the next reference point, then use your improved sketch to create the next reference point.
Thanks, Justice! Any idea how much this will cost when actually released?
They have prices up now. I think $24-$30 a month for unlimited images/videos etc
so mental gymnastics tutorial
People need to stop calling everything they make "art".
Nothing you've shown in your video is art, it's just illustration.
Almost no-one in the RUclips "art" community is making art. All these videos about AI replacing artists are so stupid because if AI can replace you, you're not an artist.
As an artist focused channel, what a big disappointment to find this garbage here. We will no longer be renewing our Tablet Pro Studio subscriptions for another year.