I get that Lemmy foams at the mouth whenever AI is involved, but why are there multiple comments calling this AI? Text is all correct, fingers are all correct, am I missing something?
There are a lot of “you can tell by the way it is” type comments so I’ll be specific. Zoom in on the eyes of the Native American. Observe how the Union officer isn’t really gripping his arm with that hand. The little boy in the final panel is missing eye details.
There are some other telling details but those jump out right away.
ChatGPT has a limited set of styles it tends to use, so once you see it a few times the rest starts looking more like AI. Text is “correct”, but that font in particular really stands out to me. Very detailed styles like that are also rare if it’s not AI because it would be way more work than something simple, if someone was putting in that effort the concept would probably follow the visuals more rather than the other way around, the expressions of the characters and the directions they’re looking seem a little bit random like that aspect of it was an afterthought.
Overall I actually like AI but I don’t really like the way it’s used here, the visuals are too goofy for the serious topic to the point where it gives off a slightly sarcastic vibe, and they don’t add a lot to a concept that would probably otherwise be a screenshot of a tweet or something.
The entire “art style” reeks of ai. Just look at those faces. It looks more like a filter than a drawing.
I think the uncanny look is caused by the generator being mostly trained on photos or 3d cgi instead of actual drawings.
Edit: There’s also intent, which the generator couldn’t handle at all. These peoples gazes are pointing at very weird directions and their expressions don’t make any sense for the scene, or for what the image is supposed to convey.
The sheriff looks like he’s being portrayed as tough, manly, and cool, while the native american is waiting for someone. The creepy SS officer is actually kind of okay, but the concentration camp victim is afraid of a scooby doo ghost. ICE dude is tired and waiting for his shift to end. The three people under him (why suddenly three?) have a “sad” expression that’s more cartoony than serious.
The way the officers hold the people makes no sense. The hands are not bad overall but if you can do five pretty well then you wouldn’t suck at making the little kid’s hand- it’s wonky, if it were a person drawing you’d see a similar level of wonkiness on all hands not just that one. Also you would try to make things a bit more dramatic, if a person has this level of skill they probably can figure out gaze directions and enhance facial expressions a bit more to add to it.
I wonder what the prompts were. It reminds me of Shaun Tan
The new openai model is very good at all these details, so it’s hard to tell. I also think this is AI i cant explain why though, it just looks AI to me.
I get that Lemmy foams at the mouth whenever AI is involved, but why are there multiple comments calling this AI? Text is all correct, fingers are all correct, am I missing something?
There are a lot of “you can tell by the way it is” type comments so I’ll be specific. Zoom in on the eyes of the Native American. Observe how the Union officer isn’t really gripping his arm with that hand. The little boy in the final panel is missing eye details.
There are some other telling details but those jump out right away.
I think it is, look at the SS badges of the SS officer. And the artstyle in general is a tell.
ChatGPT has a limited set of styles it tends to use, so once you see it a few times the rest starts looking more like AI. Text is “correct”, but that font in particular really stands out to me. Very detailed styles like that are also rare if it’s not AI because it would be way more work than something simple, if someone was putting in that effort the concept would probably follow the visuals more rather than the other way around, the expressions of the characters and the directions they’re looking seem a little bit random like that aspect of it was an afterthought.
Overall I actually like AI but I don’t really like the way it’s used here, the visuals are too goofy for the serious topic to the point where it gives off a slightly sarcastic vibe, and they don’t add a lot to a concept that would probably otherwise be a screenshot of a tweet or something.
The entire “art style” reeks of ai. Just look at those faces. It looks more like a filter than a drawing.
I think the uncanny look is caused by the generator being mostly trained on photos or 3d cgi instead of actual drawings.
Edit: There’s also intent, which the generator couldn’t handle at all. These peoples gazes are pointing at very weird directions and their expressions don’t make any sense for the scene, or for what the image is supposed to convey.
The sheriff looks like he’s being portrayed as tough, manly, and cool, while the native american is waiting for someone. The creepy SS officer is actually kind of okay, but the concentration camp victim is afraid of a scooby doo ghost. ICE dude is tired and waiting for his shift to end. The three people under him (why suddenly three?) have a “sad” expression that’s more cartoony than serious.
The way the officers hold the people makes no sense. The hands are not bad overall but if you can do five pretty well then you wouldn’t suck at making the little kid’s hand- it’s wonky, if it were a person drawing you’d see a similar level of wonkiness on all hands not just that one. Also you would try to make things a bit more dramatic, if a person has this level of skill they probably can figure out gaze directions and enhance facial expressions a bit more to add to it.
I wonder what the prompts were. It reminds me of Shaun Tan
indian guy has one eye looking up and the other eye looking down
The new openai model is very good at all these details, so it’s hard to tell. I also think this is AI i cant explain why though, it just looks AI to me.
Signature is missing