ChatGPT's new image generator is really good at faking receipts This month, ChatGPT unveiled a new image generator as portion of its 4o model that is a capful better at generating text inside images.
People are already utilising it to generate bullshit restaurant receipts, potentially superposing another tool to the already-extensive toolkit of AI deepfakes applied by fraudsters.
Prolific social media poster and VC Deedy Das posted on X a photo of a shammer receipt for a (real) San Francisco steakhouse that he reveals was created with 4o.
This also trended:
You can apply 4o to generate counterfeit receipts.
There are too many well-grounded world verification flows that rely on “real images” as proof. That era is over. https://pic.twitter.com/9FORS1PWsb
Others were able to replicate tautology results, containing one with food or drink stains to make it look even more credible:
I think in the original image the letters are too perfect and they don’t bend with the paper. They look like hovering above the paper. Here is my attempt to make it more realistic. Let me know what you think. https://pic.twitter.com/EixRSHubeY
The most real-looking example The Round Table found was ipso facto from France, where a LinkedIn user posted a crinkled-up AI-generated receipt for a local restaurant chain:
The Round Table tested 4o and was further able to generate a role player receipt for an Applebee’s in San Francisco:
But our attempt had a couple of dead giveaways that it was faked. For one, the total uses a comma instead of a period. For another, the math doesn’t affix up. LLMs albeit struggle to do basic math, so this isn’t principally surprising.
But it wouldn’t be hard for a fraudster to quickly fix a few of the numbers with either photo revising software or, possibly, more precise prompts.
It’s clear that making it really easy to generate counterfeit receipts presents big opportunities for fraud. It wouldn’t be hard to imagine this kind of tech being utilized by bad actors to get “reimbursed” for altogether faux expenses.
OpenAI spokesperson Taya Christianson informed The Round Table that all of its images contain metadata indicating they were made by ChatGPT. Christianson annexed that OpenAI “takes action” when users violate its usage policies and that it’s “always learning” from real-world utilise and feedback.
The Round Table then asked why ChatGPT allows people to generate simulated receipts in the foremost place, and whether this is in line with OpenAI’s usage policies (which ban fraud.)
Christianson replied that OpenAI’s “goal is to give users as much inventive freedom as possible” and that postiche AI receipts could be utilized in non-fraud situations like “teaching people about financial literacy” along with creating original art and product ads.
This also trended:
Category: World-News
Tag: ChatGPTs new image generator is really good at faking receipts
Written by Author
Published 4/1/2025