The Slopfather's Field Guide: Five Things to Look at Before You Hit Share
You do not need to be a tech person. You need thirty seconds and this list.
There is a photograph circulating on Facebook right now of a flooded neighborhood. A child. A wet dog. Thousands of comments saying “praying.” The flood did not happen. The child does not exist. The dog was never wet.
Someone made it in about four seconds. Someone else shared it. Then someone else shared it. By the time you saw it, it had already passed through enough hands to feel like news.
This guide is for you. Not the tech person. Not the cybersecurity expert. You. The person who uses the internet every day and does not want to be the one who accidentally spreads something fake.
You do not need special tools. You do not need to understand how AI works. You need five things to look at. That is it.
Tell #1: Look at the Hands
AI is extraordinarily good at faces. It has seen billions of photographs of human faces and it has gotten very good at making faces look real.
Hands are a different story.
Hands are complicated. Every hand is slightly different. The way fingers bend, the way knuckles sit, the way a fist closes — these things vary endlessly, and AI has not figured them out the way it has figured out eyes and noses and smiles.
So before you share an image of a person, look at the hands.
Count the fingers. Look at how they sit. Ask whether a real hand could actually be in that position.
If the hand has six fingers and the person is not a cartoon character, someone generated that image. If the fingers look melted into each other, someone generated that image. If the hand is gripping something in a way no hand has ever gripped anything, someone generated that image.
The face will lie to you. The hands almost always tell the truth.
Tell #2: Read the Signs
AI knows what signs look like. It does not know what signs say.
Think about that for a moment. If you asked AI to generate a photo of a busy coffee shop in Paris, it would give you golden light and wooden tables and a chalkboard menu on the wall. The chalkboard would look exactly like a chalkboard menu.
But if you read what the chalkboard says, it might say something like: “ERRESS & COFFEE PUBIR.”
That is not a language. That is AI doing its best impression of what letters arranged on a sign look like, without understanding that letters are supposed to mean something.
This works on everything. Storefront windows. Newspaper headlines. Book covers. Name tags. Product labels. Menu boards.
The rule is simple: zoom in and read the text. If the words are nonsense — not a foreign language you do not recognize, but actual nonsense — the image is almost certainly AI-generated.
Real photographs have real words in them. AI-generated images frequently do not.
Tell #3: Ask Why This Exists
This is the most important one. The others are about catching fakes. This one is about catching the ones that want something from you.
Every piece of content on the internet was made by someone with a reason.
Before you share something, ask: what is this asking me to do?
Sometimes the answer is obvious. Like the page. Follow the account. Share this to raise awareness. Donate to this cause. Sometimes the answer is invisible, which is actually more concerning. An image that just wants you to feel something — outrage, sadness, inspiration, fear — and share it is generating what the people who study this call “engagement.” Engagement means money. The more you react and share, the more money the account makes from ads and partnerships.
Ask yourself: does this image want me to do something?
If the answer is yes, slow down. Ask who made it and why. Look for a source. If there is no source, that is information.
The fake flood with the crying child wanted you to comment “praying.” The inspirational quote attributed to no one wanted you to follow the account. The veterans’ page with AI-generated photos of wounded soldiers wanted you to share, so that more people would see the ads running alongside it.
They all had a reason. Knowing that they have a reason is the first step to not doing what they want.
Tell #4: Check the Background
AI spends most of its effort on whatever the prompt asked for. If someone asked for a photo of a woman at a dinner party, AI will put a lot of work into the woman and the table and the food.
The background gets less attention.
Look at the edges of AI images. Look at what is happening behind the main subject. Walls that warp. Furniture that fades into nothing. Windows that look out onto a blurry smear that used to be a city. Shelves where the books have no titles. Staircases that go nowhere. Doorways that do not quite connect to the floor.
The center of the image is often convincing. The edges are where AI runs out of instructions and starts making things up.
This is especially useful for indoor scenes: kitchens, living rooms, offices, restaurants. The table will look perfect. Walk your eye around the edges of the room. Something will be wrong.
Tell #5: Does the Story Check Out?
AI can generate images. It cannot generate reality.
This tell is the simplest one: if an image is claiming to show something real, ask whether that thing actually happened.
This does not require any technology. It requires one question.
If someone shares a photograph of a celebrity at an event, that event was either a real event that was covered somewhere, or it was not. If you search the celebrity’s name and that event, and nothing comes up, the photograph probably did not come from that event.
If someone shares a news photograph of a natural disaster, that disaster either happened and is being covered by news outlets, or it did not. If you search for that disaster and the only results pointing to it are the viral post, something is wrong.
Most people skip this step because the image looks real and the caption says it is real and fifty people already shared it. Fifty people sharing something is not evidence that it is true. Fifty people may have all made the same mistake. They were trusting each other. Nobody checked.
You can check. It takes about thirty seconds and a search.
The Five Tells, All in One Place
Hands. Count the fingers. Look at how they sit. Hands are where AI falls apart.
Text. Zoom into any signs, labels, or headlines. If the words are nonsense, the image is fake.
Why does this exist? Every piece of content wants something from you. Know what it wants before you give it.
Background. The main subject is convincing. The edges are where AI runs out of instructions. Look at the edges.
Does the story check out? If the image is claiming to show something real, thirty seconds of searching will usually tell you whether it happened.
That is the whole guide.
You do not need to know anything about AI. You do not need any apps. You do not need to be young or technical or fluent in anything digital.
You need thirty seconds and a habit of asking one question before you hit share: does this make sense?
Most of the time, if you look, it does not.
The Sloppies exist because looking matters. Nominations for March are open at thesloppies.com. You have already seen slop today. Send it in.