- False receipts are now easy to generate with the Chatppt
- They are convincing and additional adjustments make them seem very real
- Openai says that metadata and use policies will protect us
The generative AI seems to solve one of the most obvious problems with the imagery generated by AI: the text. And although this is good news for anyone who was tired of seeing images containing absolute gibberish, this can also have unforeseen consequences – it turns out that Chatgpt is now good enough to create false receipts.
The counterfeits are not completely convincing if you know what to look for: they are a little too clean, a little too CGI, and sometimes the erroneous sums (as in our example above). But they are also easy to miss – and with a little violin, you can make them more realistic with folds and food stains.
For the moment, it seems that the most effective way to do so is to get a ChatPPPT to recreate an existing receipt, which can then be changed to show the prices you want.
Openai says it is not too worried, but it seems to be a new effective tool for digital fraudsters and troublemakers.
Why Openai agrees with counterfeits, for the moment
You can use 4O to generate false receipts. There are too many real world verification flows based on “real images” as proof. This time is over. pic.twitter.com/9fors1pwsbMarch 29, 2025
We contacted Openai and a spokesperson told us that he had an eye on trends like the receipts generated by AI. “We monitor the generations of images on and out of our platform, use internal tools to verify that they have been created by our products and take measures when we identify violations of our use policies,” they told us.
“We always learn from the use and comments of the real world, and we will continue to refine our policies to balance the freedom of creation by preventing improper use. All the images include the standard C2PA metadata of the industry indicating that they were generated by AIA by OPENAI,” added the spokesperson of Openai.
However, Openai’s spokesperson Taya Christianson went a little further, telling Techcrunch that there was a positive potential use for false receipts: “Teaching people financial literacy” in non-fraudulent situations.
I can also imagine using it to illustrate articles or to create accessories for real or virtual staging. But false receipts can also be used for fraud, the more the images are convincing, the more they are of fraud to facilitate.
These guarantees of metadata “manufactured by Chatgpt” are very good, but exporting an image without metadata is a work with one click. I am not sure that these guarantees will do the slightest thing to prevent false images from being used more for fraud than social media “be nice” have prevented people from being horrible on the Internet.
But I must admit that it would be very funny if the OPENAI’s own employees were starting to use the technology to inflate their spending complaints.
