Recely, ChatGPT has iroduced a new image production capability as part of the 4O model that works very well in the production of the text inside the images and can create natural fake receipts. Differe people use this capability to create fake restaura receipts, which can add another tool to the extensive set of artificial ielligence technologies used by fraudsters.
Investor and social networking user, Dawa Das, posted a picture of a fake receipt of a real steak in San Francisco, which he said was created by the 4O model.
Other users have also been able to reproduce similar results, including receipts that had food stains or drinks to make them look more realistic.
Use ChatGpt to produce fake receipts
In France, a user in LinkedIn released an image of a fake wrinkled receipt created with artificial ielligence for a chain restaura that looks very natural.
It is clear that the opportunity to create fake receipts simply provides great opportunities for fraud. It is not difficult to imagine the use of this technology by beneficiaries to get fake expenses repay.
According to Openai spokesman Tia Cristinson, all the images created by ChatPT include metadite, indicating that these images were made using this tool. He emphasized that OpenIA will take the necessary steps if you see a violation of users’ use policies and always uses real feedback and use to improve the performance and increase the security of its models.
However, the question is why ChatGPT esseially enables fake receipts, and whether it coradicts Openai’s use policies that have explicitly banned fraud.
Christinson says the Open -AI goal is to deliver the highest creative freedom to users. He also poied out that fake artificial ielligence receipts are not necessarily used for fraudule purposes, but can be used in various areas such as financial literacy education, original artwork, advertising design and other creative applications.
