This month, ChatGPT unveiled a brand new picture generator as a part of its 4o mannequin that could be a lot higher at producing textual content inside photographs.
Individuals are already utilizing it to generate pretend restaurant receipts, probably including one other software to the already-extensive toolkit of AI deepfakes utilized by fraudsters.
Prolific social media poster and VC Deedy Das posted on X a photograph of a pretend receipt for a (actual) San Francisco steakhouse that he says was created with 4o.
Others have been capable of replicate comparable outcomes, together with one with meals or drink stains to make it look much more genuine:
Essentially the most real-looking instance TechCrunch discovered was really from France, the place a LinkedIn consumer posted a crinkled-up AI-generated receipt for a neighborhood restaurant chain:
TechCrunch examined 4o and was additionally capable of generate a pretend receipt for an Applebee’s in San Francisco:
However our try had a few useless giveaways that it was faked. For one, the overall makes use of a comma as an alternative of a interval. For one more, the maths doesn’t add up. LLMs nonetheless battle to do primary math, so this isn’t significantly stunning.
However it wouldn’t be arduous for a fraudster to rapidly repair a number of of the numbers with both picture enhancing software program or, presumably, extra exact prompts.
It’s clear that making it very easy to generate pretend receipts presents large alternatives for fraud. It wouldn’t be arduous to think about this type of tech being utilized by unhealthy actors to get “reimbursed” for totally pretend bills.
OpenAI spokesperson Taya Christianson advised TechCrunch that every one of its photographs embrace metadata indicating they have been made by ChatGPT. Christianson added that OpenAI “takes motion” when customers violate its utilization insurance policies and that it’s “all the time studying” from real-world use and suggestions.
TechCrunch then requested why ChatGPT permits individuals to generate pretend receipts within the first place, and whether or not that is in step with OpenAI’s utilization insurance policies (which ban fraud.)
Christianson replied that OpenAI’s “objective is to offer customers as a lot inventive freedom as potential” and that pretend AI receipts could possibly be utilized in non-fraud conditions like “instructing individuals about monetary literacy” together with creating unique artwork and product advertisements.
