Share your AI art

  • Thread starter RYAN
  • 206 comments
  • 18,847 views
I've been having too much fun with Bing's Image Generator
_9b89a58f-ad4f-4c57-a057-f5808e981d28.jpg
_3839c07e-0ca6-4732-912a-13f2253cf016.jpg
_7917bbb9-ca01-466c-a766-c3af4f975246.jpg
_f8982f98-50cf-4a61-9ea6-416c74b76592.jpg
_81921f20-9233-4401-a253-b96d4a07a004.jpg
_3352acfd-874c-4aeb-9b6f-e8ea8c8f0c38.jpg
_eaa4a1c2-f0b4-4050-9531-7efa3b623894.jpg
_1f38454f-6b4b-42c7-bba7-18e9fb94bfe6.jpg
_c93519a7-9634-49c4-8606-3adef1324d60.jpg

The thing that impresses me are the hands and feet, which are done EXACTLY as I wanted them to. They are just purrfect.

A little while back I started working on a story with dragons (nothing came of it) but I did have descriptions for all of them. Sadly I wasn't able to find all the info and the piece of writing I have on my computer didn't help, but I tried.

Anyway, this one is Barn Stormer
_5b49a994-ba18-45c3-99ba-17f3ba083b1e.jpg

This one is Hell Bringer
_b2da9a70-7d11-44a6-8b78-3f0566fe6140.jpg

And this one is Mind Killer
_c22ecc1d-c173-43a1-a501-52802f35764b.jpg

I also created the character of Project Shadowhunt, which I envision as a black lizard/human/alien thing with psychic powers. The results I got were soo much better than I expected.
_a30c2add-42cc-4c4d-b433-469ef227b3a4.jpg
_ecf369c9-148e-4c83-b487-9e5a23594daf.jpg
_e32df0aa-8411-42f2-a095-1e794c110998.jpg
_c7fc9bc7-ec89-4823-8d50-fbea0b6033e7.jpg
_9b394f3c-400a-4c44-bf97-7f9dc9a4f88e.jpg
 
Last edited:
I've been using a combination of Bing's Image generator and Adobe Photoshop's AI tools to create some INSANE stuff. Namely some characters I made for a story titled Project Shadowhunt. This one here is supposed to be "Mystic A" - a Mew meets lizard meets human child creature that has physic powers.
Original Image:
_7cc0b393-54db-485a-9884-4cf9d19c7b7e.jpg

Extended with Photoshop:
_7cc0b393-54db-485a-9884-4cf9d19c7b7e extended copy.jpg

I also have a character named "Mystic B" which is essentially the "Mewtwo" variant of Mystic A. Bigger, stronger and a whole lot deadlier, he's a genetic experiment gone wrong.

Bing Image (I didn't edit this in Photoshop)
_87732545-76ad-457c-bc5f-6f7c488f254d.jpg

53323477756_06c3928a13_o.jpg

This here is a Bing Image:
53323821989_d4d4dd10bc_o.jpg

Extended with Photoshop
53323821989_d4d4dd10bc_o extended copy.jpg

I LOVE the level of SASS contained in this image. Like, come on man, is that all you got?
To which Project Shadowhunt replies, Yes.

I don't know if it was intentional but Bing basically gave me photos of Shadowhunt's growth stages as he's made in a lab.
Bing Image
_5f2cdb37-8cf1-4a1d-aa01-11aae1370872.jpg

Extended with Photoshop
_5f2cdb37-8cf1-4a1d-aa01-11aae1370872 extended copy.jpg

Bing Image
_57a64dad-552f-4772-afa2-200e312bf828.jpg

Bing Image (Not really sure what happened here. Might make some feel a bit queezy)
_9fe40bf3-a1f3-4a58-92ab-6d508933e298.jpg
Bing Image
_9671fd93-4d2b-4568-9a59-d7386f77df3b.jpg

Bing Image
53323704303_27a98cb433_o.jpg

Literally a shadow of Mystic B.
 
Last edited:
For fans of the Assassins' Creed series, here's a slide show I posted to YouTube of AC-inspired art.

(Via ChatGPT - 4 and DALL-E 3)

Edit: These images took about an hour to create.



A sample...

1700424330629.jpeg
 
Last edited:
Here's the prompt -

"Image. Black background. A solid gold crested bird which has the shape of a cardinal, and a cobra are fighting each other around a circle of gold. The bird is gold, as are its feathers, legs, claws are gold as is the entire snake and its scales, they are aggressive towards each other, with neither dominating. Clouds of coarse gold flakes are flying off the feathers, indicative of the ferocity of the conflict."

The prompt was the result of a long evolution of ideas.

1700922851178.jpeg
 
IRON LYNX
_be929356-688a-42a5-90a3-15a977ede892.jpg

_54e5e3a7-54b7-4e3c-9686-10cbe269500e.jpg

_fb614533-4bab-4dab-b5ff-70c87b7d1cf7.jpg
_ce8afabf-d989-4931-acdb-ed96e537a80c.jpg

I think the art directors working on the Sonic 3 movie might be out of a job:
_aa930a36-57e4-4184-8ed1-0d8691ba0dd3.jpg

_4cf969c9-a1fa-4aa1-9c7c-1f67977f5713.jpg

_a17abc74-7000-4c02-82ad-14cc53c99f59.jpg

_8bd1b3d9-710a-4945-af18-85006e8e45f6.jpg
Why does the AI have to go soo hard?

I also got the AI to make a more realistic Puss In Boots:
_6c5ad2f8-6850-4462-848e-4c959958bef7.jpg
_20bce674-5aa2-4278-af8a-85cbf3c85cc9.jpg
_09ba69cb-8cb0-4252-919c-e528a3ab8380.jpg
 
Last edited:
Just messing around with generating some PS2-aesthetic Race Queens, hopefully these aren't too spicy for GTPlanet...

(Made using a local Stable Diffusion setup, but it's been over a month since I made these so I don't remember which model(s) I specifically used)
 

Attachments

  • 008.png
    008.png
    453.7 KB · Views: 15
  • 007.png
    007.png
    479.2 KB · Views: 14
  • 006.png
    006.png
    401.7 KB · Views: 12
  • 005.png
    005.png
    432.8 KB · Views: 12
  • 004.png
    004.png
    368.7 KB · Views: 13
  • 003.png
    003.png
    383.7 KB · Views: 14
  • 002.png
    002.png
    440.1 KB · Views: 14
  • 001.png
    001.png
    348.5 KB · Views: 15
Last edited:
Recently there's been talks about how dangling an imaginary conditional "tip" in front of ChatGPT can actually result in improved output, so out of curiosity I tested out something similar with Stable Diffusion.

The following were all generated using the same model (the SDXL-based Omnium v1.1), seed 0, 25 steps, CFG 3.0, DPM++ SDE Karras, and the same simple prompt of "painting of a dragon," ...but with different "commission" prices added to the prompt.

Here's "painting of a dragon, $5 commission":
painting_of_a_dragon___5_commission_0.png


Here's "painting of a dragon, $50 commission":
painting_of_a_dragon___50_commission_0.png


And here's "painting of a dragon, $500 commission":
painting_of_a_dragon___500_commission_0.png


The differences are minor and potentially up to personal preference as to which is "better"... but for giggles, I decided to try one more. Here's "painting of a dragon, $5000000 commission":
painting_of_a_dragon___5000000_commission_0.png
 
So... a very silly idea popped into my head recently, and I've decided to make it a reality: Create custom AI boxart for my Steam library. The catch is that I cannot describe the game at all, I can only use the specific prompt format of "Steam boxart for a videogame titled [GAME NAME]"

Additionally, I can only generate three images per game to choose from.

I've only done a small handful of my library so far (I've got well over a thousand games in total, so it's gonna take a while to say the least), but here's what I've got so far:
SteamLib.jpg
 
Last edited:
It being January 6th today, I asked ChatGPT to create an image to memorialize the terrible event. It refused:-

I'm unable to create or provide images related to real-world violence or sensitive historical events. If you have any other topics or themes you'd like to explore, feel free to let me know, and I'd be happy to help create or discuss those!

The next prompt was - Do it for a work of fiction

1704549158107.jpeg
 
If Costco made a GT3 Race car
 

Attachments

  • _b1b900d2-e5d2-44c2-929f-7427ac8a56f5.jpg
    _b1b900d2-e5d2-44c2-929f-7427ac8a56f5.jpg
    149.5 KB · Views: 10
  • _279cc556-eca8-4c1a-b681-7dc559b29de8.jpg
    _279cc556-eca8-4c1a-b681-7dc559b29de8.jpg
    140.7 KB · Views: 10
  • _0dc76949-caea-4e4e-8bb0-d8ade521915d.jpg
    _0dc76949-caea-4e4e-8bb0-d8ade521915d.jpg
    145.9 KB · Views: 10
  • _9c441fa7-a11b-428f-969f-a66f56532ae5.jpg
    _9c441fa7-a11b-428f-969f-a66f56532ae5.jpg
    142.2 KB · Views: 8
I'm playing through Resident Evil Code Veronica right now, and a thought entered my mind... as far as I'm aware, it's the only classic style RE with fixed camera angles and tank controls to use fully polygonal 3D graphics instead of using prerendered backgrounds.

So... I was curious to see what I could do with Stable Diffusion to try and make it look like it used pre-rendered backgrounds of a similar quality to the Resident Evil remake and RE Zero.

original.png


013.png
 
Really like AI art. I’m a former 3D Artis and still love to create. Or noweday, give the AI some info so it can create.
 

Attachments

  • IMG_1022.jpeg
    IMG_1022.jpeg
    135.2 KB · Views: 5
  • IMG_1045.jpeg
    IMG_1045.jpeg
    188.8 KB · Views: 4
  • IMG_1054.jpeg
    IMG_1054.jpeg
    193.3 KB · Views: 4
  • IMG_1069.jpeg
    IMG_1069.jpeg
    241.2 KB · Views: 5
  • IMG_1105.jpeg
    IMG_1105.jpeg
    156.1 KB · Views: 5
  • IMG_1135.jpeg
    IMG_1135.jpeg
    195.5 KB · Views: 4
  • IMG_1166.jpeg
    IMG_1166.jpeg
    200.5 KB · Views: 4
  • IMG_1179.jpeg
    IMG_1179.jpeg
    183.7 KB · Views: 4
  • IMG_1260.jpeg
    IMG_1260.jpeg
    133.9 KB · Views: 4
  • IMG_1291.jpeg
    IMG_1291.jpeg
    178.9 KB · Views: 4
  • IMG_1331.jpeg
    IMG_1331.jpeg
    224.9 KB · Views: 4
  • IMG_1369.jpeg
    IMG_1369.jpeg
    203.7 KB · Views: 4
  • IMG_1389.jpeg
    IMG_1389.jpeg
    169.7 KB · Views: 4
  • IMG_1008.jpeg
    IMG_1008.jpeg
    177 KB · Views: 5
Back