I spy, with my little eye, a photo that was faked by an AI image generator! Can you spot the clues?

FB is turning into a parade of fake AI images churned out by click-farming pages. More misinformation is on the way. Learn some tricks for spotting AI photos!

y'all PLEASE spread this around. @busybussinbee and I were just talking about studying images to see if they're ai or not.

Educate thyselves my friends and let's fight ai together!!

Couple of things from an AI artist.

First, these are all good suggestions, but they are only valid for maybe the next 12 months. The ones that are timeless are looking out for "to good to be true" and checking sources and looking for outside confirmation. But the tools are only getting better.

As a demonstration, this is the same prompt on Midjourney v1 (2022) and V6 (2024):

Soon, they won't be able to be distinguished from the old fashioned photoshop-and-or-miscaptioning clickbait from the olden days.

Second, and I cannot stress this enough, check the tags. Go to the OP's site/blog/whatever. Because I run a blog called "Deepdream Nights" named after one of the earliest AI art processes, tag my work, and often include the frickin' prompts and tools used.

And I still get wanna-be Scoobies trying to yank off my mask for posts where there's a full tutorial with in progress pics under the fold.

Third, do not automatically associate a lack of AI with trustworthiness. I'm not going to go into the myriad proper uses for genAI, I know that's not likely to get traction in this thread, but if you're spending all your time paranoid about genAI stuff, you're going to be tempted to see a lack of AI as a sign of being legit on its surface, and analogue trickery still works.

You should apply this skepticism and critical thinking to ANY kind of post, AI or not.

Making your bullshit detector work universally will protect you from advancements in the tech.

Also, photoshop still works on AI images. If the creator spends 45 minutes in photo-editing, most of your tells can be eliminated.