After the recent update of GPT 4.o, Dall-E has been left behind, because the aforementioned model has evolved to the point where can generate a cup of full winesomething I couldn’t do before.
But that’s not all, now there is a new tendency to convert real photos into illustrations similar to the style of Studio Ghibli. This has definitely marked a before and after in the opening tool, since currently It is the most powerful compared to other chatbotslike Microsoft Copilot Google Gemini.
However, the possibility of replacing humans and even Hayao Miyazaki does not agree with this technology revolution. Why does a problem represent? Not only is it going to generate unemployment in the future, but also involves many risks.
The scammers and hackers could be taking advantage of the potential of artificial intelligence and the recent event of The creation of photos of invoices with GPT-4.O It is an example of this. This is already happening and things do not paint well.
GPT-4.O is capable of generating such realistic receipts that could serve as a scam
Deedy that
Yes, the image you are seeing at this time has been generated with artificial intelligence, the same they are using to create Ghibli style photos. With the naked eye It can go unnoticed as a normal photo They send you by a WhatsApp or Telegram chat.
This sample has been shared by the influencer Deedy das in X, who has shown that it is possible Make an invoice of a real restaurant using the appropriate prompt. According to what he says, it is possible to share something like this for the “verification flows in the real world that depend on ‘real images’ as proof”.
To its publication, another outstanding figure of the social network responded, who has taken as a reference its image generated to produce another similar one, but with more realistic touches, such as Mild wrinkles on paper and food stains.
Raphael Chenol
However, one of the most impressive of all in this stir has been that of a LinkedIn user named Raphael Chenol, which has confirmed that falsified falsified documents with chatgpt can be producedexplaining that «when we spend time talking about ethics in the context of AI, it is important to denounce possible abuse.»
The images you have shared could be confused without any problem with a realbecause there are many details and imperfections that make them see true. Apparently focuses on a place called “Hippopotamus”where wrinkles and print ink are very close to reality.
In turn, he has shown an example of how this type of photographs in food or online stores that request a proof of payment for this method could be used.
Is it possible that a business can fall into the trap of invoices with AI?
Generated with Ia
Apparently, there are possibilities that Sam Altman’s great capacity is exploited by cybercriminals in several ways. As part of a plan very elaborated with voice impersonations (Deepfake), posts or pHishing messages and fraudulent links, they could convince the victim to provide banking data.
However, it may be uncommon, at least today. The AI is still susceptible to making mistakes in certain accountswords or numbers. Similarly, billing with QR code cryptograms, Blockchain verification or advanced OCR integration are useful to avoid being deceived by artificial intelligence photos.
Be that as it may, this It is a warning for all CFO and HR of companiessince not having the mentioned methods, you have to redouble the forces. According to Techcrunch, they have managed to talk to Taya Christianson, a worker of the Openia communication team, who mentioned that they «take measures» to avoid it.
In turn, he said that The main mission of the company is that there is wide creative freedom As for the generation of images and that these invoices or other documents created by the platform could be used in the field of education.
Therefore, it is most likely that the model incorporates new policies for use over time, since the company works to avoid these possible situations.
Know How we work in NoticiasVE.
Tags: Artificial Intelligence, Design, Scam