you are viewing a single comment's thread
view the rest of the comments
[+] -8 points 22 hours ago* (last edited 22 hours ago) (5 children)

First of all, they did do a bad job with their blog post.


…But I’m also shocked by the number of people who think they’re human AI detectors.

I suspect people are a year or two behind, looking for signs of messy Stable Diffusion XL output, like suspiciously styled appendages or that weird framing 1.5 always did. Or maybe the Ghibli-esque style Tweeters used to ape Sam Altman.

Models aren’t like that any more.

The other day, I was sitting at a TV, looking at photos side by side, I cannot tell my own mirrorless camera RAWs apart from some locally-run generations that used my photos as a reference, even if I zoom in to pixel peep or try to nitpick the depth-of-field from my lens.

Videos edited by H3 are shockingly good now, albeit at lower resolution.

I did some animesque character design mockups (just for my own thinking/brainstorming), and I can feed the model separate reference images for characters and art styles and poses and it… just gets it. It looks just like the original painted style, and I can’t find any artifacts or distortions that jump out. Not counting the ones in the original material.


I’m not trying to glaze diffusion or anything; quite the opposite. It’s messy, and sloppy. Besides, that’s not the point, and I don’t want to get into that.

What I’m saying, outside of really lazy slop or bad models like ChatGPT, people are behind if they think they can spot AI-generated stuff reliably.


So, maybe the studio lied.

The blog post certainly makes it suspicious. They could have uploaded some asset at least?

But, as suspicious as the style is, I can’t tell if that “process” picture is AI generated. Certainly not because the 2nd frame has a cartoon style, or the 3rd and 4th look like weird cg.

  • source
  • hideshow 10 child comments
  • [–] 3 points 18 hours ago

    What I think the average person who may not have an eye for art or an attention to details might not understand is, a lot of AI art that comes across as convincing to the untrained eye is still, at the very least, living deep in the uncanny valley.

    Much in the same way that an AI voice or video can look somewhat convincing but a little bit off, the same thing is true for AI artwork. People who are artists or who have at least learned to see the nuance in art can often see the weirdness in generative AI "art", even if it is sometimes hard to explain exactly why it feels off (maybe it's the subtlety of the lighting, the shape design, the line quality, etc.).

    Can you make AI art that is convincing enough to fool an untrained eye? Easily.

    Can you make AI art that is convincing enough to feel a trained eye? Sure, but it's going to be a lot harder.

    In the case of these guys, they really aren't fooling almost anyone with their obvious slop. Frankly, it doesn't take a well-tuned eye for artistry to see that their artwork looks like every generic AI slop art output that we've been barraged with over the last year and change.

  • source
  • parent
  • [–] 5 points 21 hours ago (1 child)

    I'm not going to claim to be good at detecting AI generated content, but for visual content there's clear vibes that AI generated content give that I can never specifically point out, but when I spot them there's usually at least one other person calling out the slop. Yes 3-4 years ago diffusion models would make errors like mangling hands and mushy details, but while newer models get images technically correct, they leave clear vibes that are easy to spot

  • source
  • parent
  • hideshow 2 child comments
  • [–] 2 points 19 hours ago

    Motion errors are also still very much spottable on AI video. Shit blurs, but not in the way caused by actual motion being captured on film and causes distortion of the thing moving in ways that don't happen with motion blur (like a waving arm bending at points anywhere other than the wrist and elbow).

  • source
  • parent
  • [–] 2 points 20 hours ago

    I fully admit that there are AI-generated images that could fool me. But that's beside the point when like in this case, the images are in a very typical AI-style. It doesn't look like they've put any effort into generating something that could fool people, but just went with what one of the more common models shat out. And those are still fairly obvious to detect IMO. Besides that, their process pictures really don't make a lot of logical sense. I also find it quite unlikely that a human artist would miss a visual gag like the cow using a milk churn and just replace it with a generic blaster.

  • source
  • parent
  • [–] -3 points 22 hours ago (1 child)

    Guy, everyone thinks they're a Human AI Detector, especially if they have an account on Twitter, Reddit, BlueSky, etc. Which often results in baseless accusations against human made content being called AI generated, which is hilarious to me honestly.

  • source
  • parent
  • hideshow 2 child comments
  • [–] -2 points 22 hours ago* (1 child)

    Well, apparently Nintendo fans and commenters in the linked news section do too.

    But I don’t want do disqualify how sus the blog post is, either. It does feel a little like a corporate paint-over; the internet can reasonably be suspicious of that. But I haven’t been convinced by a single comment trying to dissect the picture.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 2 points 18 hours ago* (last edited 18 hours ago)

    I don't know anything about this particular situation so won't comment on it either way. I just think its funny that so many people can be so confidently wrong when it comes to thinking something is AI generated.

    I get when something feels suspicious, and often feel this myself, but to be so confident to say it as fact and then be totally wrong when the artist posts a timelapse or something is hilarious to me.

  • source
  • parent