A social media user has caused a wave of alarm on the Internet. A user known as Shirish on X shared a hyper-realistic check generated by AI.
It appeared to be an official UCO bank check ₹69,000. The check was payable in Shirish’s own name. It claimed to have been created using OpenAI’s newly launched ChatGPT Images 2.0 model. His post carried the chilling headline: “We are so cooked.”
Read also | Trump deleted a post with an image of an AI resembling Jesus after a backlash from Christian voices
The image went viral almost immediately. It showed every element of a real bank check. The account number, bank branch details, signature line and MICR strip at the bottom looked authentic. The level of detail has alarmed thousands of users across the platform.
Reaction on social networks
Reactions on social media were divided but largely concerned. One user wrote that the pictures and videos can no longer be trusted. One user called it a serious violation of law and policy.
“This is a serious violation of law and policy. How can an AI create a fake check?” wondered the user.
Another said that ChatGPT should not have allowed these outputs in the first place. Some users remained skeptical about the real risk of fraud.
Banks require account holders to verify check details for checks over a certain amount. Without the actual barcodes and account numbers, the check cannot be cashed.
According to one user, things will get really dangerous when AI can generate functional barcodes that banks will accept. Another user said that AI has already generated a working barcode from an existing one.
Some answers were lighter.
“So you got ₹69,000?” asked one user. Another joked that the real problem was just asking ₹69,000.
“We’re done,” another user exclaimed.
Note: This report is based on user generated content from social media. LiveMint has not independently verified and does not endorse these claims.
Read also | Google launches Nano Banana 2 after viral success of AI image generation tool
ChatGPT Images 2.0: Are we ‘cooked’?
Until recently, image AI tools struggled with text inside images. Ask an AI to write words on a billboard and you’ll get garbled letters. That no longer applies.
ChatGPT Images 2.0 can now accurately render small, readable text in images. Imagine a cinema poster with the actors’ names clearly printed. Or imagine a restaurant menu with legible prices and names of dishes. Or it’s a fake boarding pass with a real looking flight number and gate.
The model can generate functional QR codes. Scan one and it actually takes you somewhere. Earlier models created decorative patterns that scanned like nothing.
Read also | What is Grok Imagine? How does it compare to ChatGPT, Gemini and others
Supports images up to 2K resolution. Fake ID now looks sharp even when zoomed in.
It handles dense layouts with precision. Imagine prescription pads, government forms, official letterheads and bank documents, all looking completely authentic.
Reads and writes correctly in languages other than English. A fake Aadhaar style document in Hindi or Japanese official letter is now within reach.
It can crawl the web in real time while generating. A poster about a real upcoming event can include exact dates and venue details.
OpenAI says the model is available to all ChatGPT users today. Thought-based image features are available to Plus, Pro and Business subscribers.
So we are “cooked”? That depends on who is using it. Making an A-bomb and dropping it on Hiroshima are two different things.
