the way you described us as not being able to recognise ai made me actually question myself. like slowing down and appreciate the world kinda question myself. lol
That's okay! I had that moment of questioning myself too. Hence why this video exists. It's not your fault. The current world we live in makes us go fast.
I actually have vision issues like needing glasses and having issues perceiving the difference between green and beige, so sometimes download cute artwork only to have someone tell me it looks off later. It's always so frustrating because it's usually things for niche fandoms that often don't get fanart. Recently I've been trying to draw these things myself, but am having trouble due to lack of authentic reference photos for that specific topic that aren't AI generated.
It IS frustrating trying to find good references now :/ I feel you. If you're interested Duchess Celestia has a very good video recently about how to spot differences in AI character art that should be okay. Though having to do it at all is exhausting.
Try typing into Google images the years 2020, 2019, anything earlier than 2021 or 2022, when dal-e and other image generators like nightcafe and midjourney started pumping out a ton of slop. Unfortunately this might not work for newer fandoms.
I feel its rather interesting to study the ai imagie itself , on what exactly produces so much likeness that causes the human brain to overlook mistakes upon first glance and such. Also i find it ez to identify usually bcuz of how much "detail" and cluter the ai tries to add to one background and the lighting
Along with the 'not looking as hard at backgrounds' idea, I think it's also worth noting that maybe one reason we pick up on AI images of people quicker is because humans are very good (usually) at telling when another image of a human looks 'wrong', even if we don't quite know why it's wrong, which is where we get the concept of the uncanny valley. Objects in backgrounds probably don't trip the same alarm bells in our brains as quickly if we're not trained to notice when they're off, so that could be why a lot of people take longer to twig that a background is off in the same way we'd notice a rendering of a person being off.
Great video with an important message. Also, you have a lovely model!^^ Edit: as someone who did backgrounds for a webtoon for a brief moment... I wish background artists would get more respect. Good background art feels natural. Kind of like music/ambiance in movies. And yeah in ai art, there are inconsistencies.
It's a kind of a bit like how you can't truly hear some parts of audio, or the bits that make up a compressed image or video. Your eyes pass over lots of things, naturally, because your brain has to process so much informatiok, is discards most of it.
AI images always, (usually) eventually give me this icky feeling and it's hard to describe. It's like nausea or vertigo. Maybe it's because I'm a millennial and I'm old enough to recognize messed up digitized objects vs real life objects and mismatches of lighting and depth, so my brain feels sick trying to make Sense of a nonsensical false photo. I also grew up playing with Photoshop and doing photo bashing/composite collage making so I have a bit of educational background in the joy of Photoshop fails. But AI images still feel uncanny and eerie, like a confused nightmare.
Ooooh I hadn't thought about the age angle, but you bring up a good point. Gen X and Millennials are the last gens to remember a world where the internet didn't dominate our lives. I could definitely see that playing a role in that uncanny valley feeling AI work gives off.
the way you described us as not being able to recognise ai made me actually question myself.
like slowing down and appreciate the world kinda question myself. lol
That's okay! I had that moment of questioning myself too. Hence why this video exists. It's not your fault. The current world we live in makes us go fast.
I remember when you found that reference! We were so salty when we found out it was AI. I'm glad you were able to take it and make into actual art! XD
Thanks a bunch!
I actually have vision issues like needing glasses and having issues perceiving the difference between green and beige, so sometimes download cute artwork only to have someone tell me it looks off later. It's always so frustrating because it's usually things for niche fandoms that often don't get fanart.
Recently I've been trying to draw these things myself, but am having trouble due to lack of authentic reference photos for that specific topic that aren't AI generated.
It IS frustrating trying to find good references now :/ I feel you. If you're interested Duchess Celestia has a very good video recently about how to spot differences in AI character art that should be okay. Though having to do it at all is exhausting.
Try typing into Google images the years 2020, 2019, anything earlier than 2021 or 2022, when dal-e and other image generators like nightcafe and midjourney started pumping out a ton of slop. Unfortunately this might not work for newer fandoms.
I feel its rather interesting to study the ai imagie itself , on what exactly produces so much likeness that causes the human brain to overlook mistakes upon first glance and such.
Also i find it ez to identify usually bcuz of how much "detail" and cluter the ai tries to add to one background and the lighting
That's actually a really good point! It can be interesting in its own way.
This video is great! Your channel is so underrated
Oh my goodness thank you so, SO much! This made my day :D
This vid is rly great to watch when paired with my current move to reduce social media in my life. Thanks!
Along with the 'not looking as hard at backgrounds' idea, I think it's also worth noting that maybe one reason we pick up on AI images of people quicker is because humans are very good (usually) at telling when another image of a human looks 'wrong', even if we don't quite know why it's wrong, which is where we get the concept of the uncanny valley. Objects in backgrounds probably don't trip the same alarm bells in our brains as quickly if we're not trained to notice when they're off, so that could be why a lot of people take longer to twig that a background is off in the same way we'd notice a rendering of a person being off.
Great video with an important message. Also, you have a lovely model!^^
Edit: as someone who did backgrounds for a webtoon for a brief moment... I wish background artists would get more respect. Good background art feels natural. Kind of like music/ambiance in movies. And yeah in ai art, there are inconsistencies.
Thank you so much! I appreciate it
Bayar has left the game to jump into your drawings. :D
He has! :D I wanna keep practicing too so I can do that more often.
I love your hair!!! And this video!!
THANK YOU! I appreciate you watching
Very interesting video! I’m not sure what I have the hardest time identifying though
That's okay! I appreciate you took the time to watch :D
It's a kind of a bit like how you can't truly hear some parts of audio, or the bits that make up a compressed image or video. Your eyes pass over lots of things, naturally, because your brain has to process so much informatiok, is discards most of it.
AI images always, (usually) eventually give me this icky feeling and it's hard to describe. It's like nausea or vertigo. Maybe it's because I'm a millennial and I'm old enough to recognize messed up digitized objects vs real life objects and mismatches of lighting and depth, so my brain feels sick trying to make Sense of a nonsensical false photo. I also grew up playing with Photoshop and doing photo bashing/composite collage making so I have a bit of educational background in the joy of Photoshop fails. But AI images still feel uncanny and eerie, like a confused nightmare.
Ooooh I hadn't thought about the age angle, but you bring up a good point. Gen X and Millennials are the last gens to remember a world where the internet didn't dominate our lives. I could definitely see that playing a role in that uncanny valley feeling AI work gives off.