'You can use AI to generate fake receipts': Viral post shows hyper-realistic restaurant bill, sparks debate
A viral post on X revealed how AI can generate hyper-realistic receipts, raising concerns about fraud and online verification systems.
A man took to X (formerly Twitter) to share an AI-generated receipt, claiming that tools like GPT-4o can create hyper-realistic fake receipts that could be used to bypass verification systems.
“You can use 4o to generate fake receipts. There are too many real-world verification flows that rely on ‘real images’ as proof. That era is over,” he wrote, sharing an image of a restaurant bill.
Also read: ChatGPT chief's appeal as Studio Ghibli breaks internet: ‘Can yall please chill’
The AI-generated receipt showed a bill from "Epic Steakhouse" in San Francisco, listing items like filet mignon, rib eye, and a Caesar salad, with a total of $277.02. The image featured wrinkles, realistic printing, and even a wooden table background—details that made it nearly indistinguishable from a real receipt.
Take a look at the post:
The post quickly sparked reactions, with one user commenting: "Generate me a photorealistic iPhone picture of a $277.02 wrinkled receipt on a wooden table with reasonable numbers. Make the math add up. The restaurant name is X and the address should be Y.”
Another pointed out an inconsistency, saying, “You can tell it’s fake by the fact there is nothing at Epic that…”
A user added, "I've already used AI image generation to that it will soon break Age Verification. Just proof of concept and a couple months back, but give it 2 years. Prompt was something like "guy in early thirties holding his passport by his face, showing the photo"."
One user commented, “n Europe, every receipt has a QR code that you can scan, which takes you to the tax authority’s website, where all the transaction details are listed.”
A user added, “I think if you do this to any degree that matters, companies will notice and you will get in big trouble. I also think that people who would be doing this to their companies are probably already doing it in other ways. So I don’t think this really moves the needle much.”
