Confuse AI? Fuck, I’m confused…wait…OMG…am I AI!!!
Absolutely human, i even removed the context text and it didn’t even flinch.
“Similar color” yep seems right, “…and texture” wait what?
idk what kind of muffin they are feeding this AI but sure is a hairy one
Cromch
Wow, that’s a little too impressive. I’m guessing that image was probably in its training set (or each individual image). There are open training sets with adversarial images, and these images may have come from them. Every time I’ve tried to use ChatGPT with images it has kinda failed (on electronic schematics, plant classification, images of complex math equations, etc). I’m kind of surprised OpenAI doesn’t just offload some tasks to purpose-built models (an OCR or a classification model like inaturalist’s would’ve performed better in some of my tests).
This exact image (without the caption-header of course) was on one of the slides for one of the machine-learning related courses at my college, so I assume it’s definitely out there somewhere and also was likely part of the training sets used by OpenAI. Also, the image in those slides has a different watermark at the bottom left, so it’s fair to assume it’s made its rounds.
Contradictory to this post, it was used as an example for a problem that machine learning can solve far better than any algorithms humans would come up with.
Just take a bite out of each. You know you got the right one when its starts yelping.
Did people make you? If so, yes!
wow!! such pretty orange houseplants
Ah yes, I grow these orange houseplants every year. Visitors definitely admire this species of lovely orange houseplant 😉
Unfortunately, AI is quickly becoming better at this kind of thing than average humans are. And the internet is doomed for sure as a result.
We’re gonna be swimming in a sea of AI generated, convincing half truths and lies.
I’m pretty sure gpt4 can do this
Yep, 0 errors. Couldn’t find a video I saw
I like the way one of those pictures is of a slightly deformed dog. Figure THESE out, AI!
(No please don’t. We don’t want AI to figure out any more than it already has.)
I only see deformed dogs in the non-muffin ones.
Are you sure those aren’t the muffins? :P
That’s because shitty people love their dogs deformed, unable to breathe, and with high vet bills. They think it’s cute or purebred.
12 is definitely abnormally shaped.
yup thatsa the one ima talkin about
Or the next Google CAPTCHA after one of the sneaker-scalpers trains their own.
I want a War on AI so bad
The cylons won’t take kindly to this…
I don’t! I want people who are concerned about the misuse of AI, particularly by corporations and world governments, to learn how to use AI to fight back against our oppressors or at least make AI-powered technologies that are helpful for common people, and to archive how it works, particularly how it fails.
I hope each of them is named muffin
Click all the images containing something sweet.
All of them.
is it worth it?
Artists are now fighting back against AI stealing their works
Is that the real war on AI?
Human capitalists fighting for their profit VS other Humans that fight for some rights of theirs.
The age of digital war is really settling in now that even protesting is something we do with digital tool.
Glaze has been defeated a long while ago.
Cookiedogs > Skynet! Quick! To the ovenkennel!
Mmmm blueberry muffins…
My favorite new captcha
Tried it on chat gpt 4, here what it thinks:
This image is a collage that alternates between photos of Chihuahuas and blueberry muffins. The arrangement is such that it plays on the visual similarities between the two, with the muffins and parts of the Chihuahuas (likely their faces) mimicking each other in color and texture. This creates a humorous effect, as it can be initially challenging to differentiate between the muffins and the dogs.
Starting from the top left corner and moving left to right, row by row:
- Blueberry muffin with spots resembling a dog’s eyes and nose.
- Face of a Chihuahua with a similar coloration to the muffin.
- Blueberry muffin resembling the face of a Chihuahua.
- Chihuahua face with a light fur color matching the muffin’s surface.
- Chihuahua face with dark eyes and nose similar to blueberry spots.
- Muffin with a pattern that resembles a Chihuahua’s facial features.
- Chihuahua with an expression and coloring that echoes the appearance of a muffin.
- Muffin with blueberries and coloring that looks like a Chihuahua’s face.
- Chihuahua with a facial expression and fur colors that mimic a muffin’s texture.
- Muffin with blueberries mimicking the eyes and nose of a Chihuahua.
- Chihuahua with features that resemble the spots on a muffin.
- Muffin resembling a Chihuahua’s face in color and texture.
- Close-up of a Chihuahua’s face with colors similar to a blueberry muffin.
- Muffin with a pattern of blueberries resembling a Chihuahua’s face.
- Chihuahua looking directly at the camera, with fur colors like a muffin.
- Two Chihuahuas close together, with their heads resembling muffin tops.
Conclusion it sometimes gets fooled :)
It’s likely seen the image before. Try randomising the image, reversing some of them, altering the gamma, and adding some noise. See how it does then.
