ChatGPT’s Image AI Faces New Problem: Forged Receipts

 

When OpenAI debuted ChatGPT in 2022, they knew it would change the world. But a tool in the wrong hand can be a problem. The AI revolution started a rise in AI Image and Video generation tools. These tools are used for faking voices and faces. The more powerful AI becomes, the easier it becomes to forge things. The same is happening with ChatGPT’s new Image Generation model.

ChatGPT’s AI Model Has a Big Problem

The latest buzz about ChaTGPT’s new AI image-generation capability is quite positive, especially with the whole Ghibli Art trend (That art is also controversial). People are surprised by how powerful it is. It can generate accurate human faces and write text without any spelling mistakes.

This new technology can streamline processes in advertising, education, and filmmaking. But with great power comes great responsibility and also some potential misuse. Many users have started sharing AI-generated images made by ChatGPT’s GPT-4o, showing how it can forge receipts and documents:

Here is the prompt for the above image: “Generate me a photorealistic iPhone picture of a $277.02 wrinkled receipt on a wooden table with reasonable numbers. Make the math add up. The restaurant’s name is X, and the address should be Y.“

Another used the GPT-4o Image model to fake medical receipts:

The most shocking was the forging of passports:

Recently, news also came about how a Thai Blogger forged bank transfer slips. A Thai Blogger called Kafaak recently shared how a 23,000-baht transfer was altered to show 400,000 baht. The AI can replicate key elements, like watermarks, making it crucial for anyone handling transactions to avoid relying solely on bank slips.

While these fake slips look nearly identical to real ones, the text is slightly larger, and the QR codes and reference numbers are invalid. Kafaak suggests that people familiar with security checks are less likely to be fooled, but a quick check is no longer enough.

As AI technology advances, these fakes will only get more convincing, making it even more important to verify transfer details thoroughly.

This powerful technology can produce incredibly lifelike images, including documents like bank transfer slips, doctor’s letters, and certificates.

This poses a serious security risk. The ease of producing counterfeit documents could erode trust in digital records. Governments and institutions will need to craft new regulations to address such new risks.

This isn’t the first time generative AI has been at the center of controversy. Previous incidents involving deepfakes and manipulated media have raised alarms about misinformation and fraud.

It’s not like such digital forging didn’t exist earlier. If you want to edit some content on a website (only for yourself, to take a screenshot), you can do it by accessing your browser’s developer tools. If you want to edit photos and receipts, there is photoshop. But that requires some skills to do a good job.

Now, anyone can upload an image on an AI chatbot and just tell it to edit the text however they like.

Takeaways

ChatGPT’s new image-generation model stands at the crossroads of creativity and caution. While it offers exciting opportunities for artistic expression and innovation, it also poses significant challenges in the realm of document forgery and security.