ChatGPT’s new image generator is really good at faking receipts

This month, Chatgpt has introduced a new image generator As part of the 4O model that is better in creating text in the pictures.

People already use to create fake restaurant receipts, potentially use to add another tool already a large set of tool EU depth used by scammers.

Prolific Social Media Poster and VC DeDy Das Sent x 4o a picture of a fake receipt for a (real) San Francisco Steakhouse, who says it was created with.

Others were able to repeat similar results, including someone with food or drink spots to look even original:

The most realistic looking sample TechCrunch was from France where a LinkedIn user stationed AI created by a whip for a local restaurant chain AI

TechCrunch 4o was tried and at the same time San Francisco was able to create a fake receipt for an Aplebee:

But there were a few dead gifts where our attempt was fake. It uses a comma instead of a total of time for one. It does not add math to someone else. Llms Still struggling to make basic mathematicsTherefore, this is not especially surprising.

However, the cheat will not be difficult to correct a few numbers in a photo regulatory program or perhaps more accurately instructions.

False revenues offer great opportunities for fraud, which really facilitates creating. It would be difficult to imagine such techniques used to “pay” for false costs of bad actors.

Openai spokeswoman told Taya Christianson Techcrunch that includes metadata prepared by Chatgept for all his pictures. Christianson added that when users violated the rules of use and added that the use and opinion of the real world is “always learning”.

TechCrunch then asked ChatGpt to create fake receipts in people primarily and whether it was open Rules of use (Fraud is prohibited.)

Christianson replied that Openai’s “goal is to give users as creative freedom to users” and fake AI revenues can be used in situations without fraud, along with the original art and product advertising.