Empty file header, but “FF” is the magic jfif\jpg type. Because jpegs are binary files you can’t reliably trust a text string watermark showing up in a hex viewer.
Online detectors vary, but say no ai, just 14% straight edges\grid-like patterns\brush circling. No known watermarks detected. No stenography hidden patterns. The bottom right corner has an clear easy-to-drop clone stamp pattern – but could just as well be normal carpet.
TinEye says 0 results. No other file sources so I guess that’s an ai lead?
As SkaveRat said it has a synthid. To my knowledge, “ai detectors” are pretty garbage, here only openai has the keys to check their watermark. It’s in the pixels. Complicated mathematical method, but reliable and hard to remove.
What made me suspicious and check was that the light “overexposure” looked very odd to me. Especially the weird comb pattern into the cats butt. That and cats just wouldn’t lay like that when there’s sun.
Confirmed. Just now went down a rabbit hole how synthID works. A synthID can also be placed into a real photo someone has taken and it goes thru a google service – like adding captions.
SynthID is a mathematical property, based on how the neural network that generated the image “sees” images.
Basically it uses knowledge of the neural net, and a password, to make the image weird in only that networks eyes. Similar idea to visual hacks, where a minor random noise pattern added to an image makes an ai think it is something entirely else.
To make an image get a synthid, you have to make an ai generate it. If a tool adds captions and gives it a synthid, then it changed the entire image and sloppified it, not jsut those captions. It’s now an ai’s impression of what the original image was.
I suppose this is possible to happen in practice, but I’ve only seen something like that with pre ai-pocalypse things. Namely I recall some ai noise filters making human voices in old videos sound ai generated. Which back then people weren’t sensitive to but now sticks out cause we had to learn to detect those tells.
it’s ai watermarked :(
What? Where? How?
Empty file header, but “FF” is the magic jfif\jpg type. Because jpegs are binary files you can’t reliably trust a text string watermark showing up in a hex viewer.
Online detectors vary, but say no ai, just 14% straight edges\grid-like patterns\brush circling. No known watermarks detected. No stenography hidden patterns. The bottom right corner has an clear easy-to-drop clone stamp pattern – but could just as well be normal carpet.
TinEye says 0 results. No other file sources so I guess that’s an ai lead?
As SkaveRat said it has a synthid. To my knowledge, “ai detectors” are pretty garbage, here only openai has the keys to check their watermark. It’s in the pixels. Complicated mathematical method, but reliable and hard to remove.
What made me suspicious and check was that the light “overexposure” looked very odd to me. Especially the weird comb pattern into the cats butt. That and cats just wouldn’t lay like that when there’s sun.
Confirmed. Just now went down a rabbit hole how synthID works. A synthID can also be placed into a real photo someone has taken and it goes thru a google service – like adding captions.
Yikes.
SynthID is a mathematical property, based on how the neural network that generated the image “sees” images. Basically it uses knowledge of the neural net, and a password, to make the image weird in only that networks eyes. Similar idea to visual hacks, where a minor random noise pattern added to an image makes an ai think it is something entirely else.
To make an image get a synthid, you have to make an ai generate it. If a tool adds captions and gives it a synthid, then it changed the entire image and sloppified it, not jsut those captions. It’s now an ai’s impression of what the original image was.
I suppose this is possible to happen in practice, but I’ve only seen something like that with pre ai-pocalypse things. Namely I recall some ai noise filters making human voices in old videos sound ai generated. Which back then people weren’t sensitive to but now sticks out cause we had to learn to detect those tells.
It has a synthID watermark