ChatgPT can now create scary real -looking receipts.
The CATGPT new feature can make pictures of receipts that have no connection with reality. Illustration. Pedro Salaverr
Openai, an artificial intelligence company, recently updated its 4O-artificial intelligence model behind ChatgPT.
Thanks to the upgrade, chatgpt is quite skilled in creating images based on the users’ prompt. Users are already have time to get excited Among other things, ChatgPT creates the style of the Studio Ghibli animation studio in the style of photos.
However, improving image creation is not an unequivocally good thing.
ChatgPT wounds are left with most inputs for illegal or other gross fermentation.
Also read: Fear of a quick payment setting will increase scams
Also read: You would probably be better than this player – still this is considered
Also read: Do you let your child use your phone or tablet? Do this first
Techcrunchin However, according to ChatgPT, you kindly create receipts of different purchases that the user may never actually have done. This opens the doors for a malicious user for a variety of scams.
According to Techcrunch, it was apparently the first to find the first investor and social media influencer DEEDY DAS. In X, he published an illustrative picture that Chatgpt had created on request.
Deedy Das’s photo receipt is supposed to be from Epic Steakhouse restaurant in San Francisco.
If the publication does not appear, see it here.
Even more credible examples can be found
Techcrunch also found a fake detonator created by chatgpt, a French Raphael Chenol has published In LinkedIn. It is more routine than Deedy Das’s receipt, which makes it look even more credible.
Techcrunch himself also tried to create a fake receipt at chatgpt. The receipt of the Applebee’s Restaurant, made by the publication, looks like a real side.
However, there are details in the false cubes that can be identified at least at the moment.
In Raphael Chenol’s fall, chatgpt has misrepresented the street where the restaurant is located. On the other hand, the prices and total amount of food purchased in Techcrunch’s decline do not match, which the publication itself points out.
Openai spokesman Taya Christianon tells Techcrunch that the company will take any breakdowns of the terms of use seriously. The company’s terms of use prohibit fraud in its products, and the creation of receipts could thus be imagined to break them.
However, Christianson points out to Techcrunch that the receipt images made by chatgpt may also have a use that is not associated with fraud. Images can be used, for example, in art, advertising or in teaching money.