Blog Details

GPT-4o, ChatGPT’s latest brainchild, has been on an image-making spree. It has made over 700 million visuals created since launch, with many inspired by Studio Ghibli. But while users are busy creating dreamy anime-style portraits, a darker trend is emerging – fake government IDs. These include Aadhar and PAN cards, which are now showing up online.

Of all the things that caught our attention, one stood out – an Aadhaar card with Elon Musk’s face on it. Yes You Read That Right!

OpenAI recently gave ChatGPT the power to generate images natively—no DALL·E required. With that came a surge of creativity and controversy. Users have already started sharing images of photorealistic Aadhaar and PAN cards, created with eerie accuracy. QR codes, card numbers, official logos—it’s all there, just not legally.

Musk and Altman on Aadhar Cards

Several social media users posted mock Aadhar cards with images of famous personalities on them, such as Elon Musk and Sam Altman. There are also reports of fake PAN cards being generated.

While AI misuse has long been a concern, GPT-4o’s ability to render such realistic images brings that concern into sharp focus. Even though OpenAI has restrictions in place, ID forgery wasn’t off the table.

How is this even possible?

GPT-4o’s autoregressive model gives it a significant edge over previous diffusion models like DALL-E. That means it can understand complex, nuanced prompts. This is great for users who want custom visuals, but also great for those with less noble intentions.

OpenAI acknowledged the risks in its system card, noting that this new model may pose “risks across a number of areas, in ways that previous models could not”.

What are the experts saying?

Wriju Ray, CBO at IDfy, said that while Aadhaar data can be validated through backend systems, PAN and driver’s licenses are trickier, since these often lack facial data in official databases. However, he further added, “Aadhaar is easier because you can match text and face data. But with PAN and DL, you need to look out for tampering,”

Rohit Kumar of The Quantum Hub added that the real challenge lies in controlling what AI outputs, not just how it’s trained. Watermarks, denial systems, and metadata tagging could help stem the flow of fakes. “If we don’t act now, we risk a societal trust crisis,” he warned.

How to Differentiate Real from Fake?

It has now become crucial to distinguish between a genuine Aadhaar card and an AI generated one. Here’s a small guide to know how.

The Mugshot Mismatch: A bona fide AAdhar card features your actual photograph. AI-generated fakes often sport images that don’t quite hit the mark, sometimes bearing no resemblance to the original.

Font Faux Pas: Authentic Aadhar cards maintain consistent Hindi and English fonts. If the text looks like it’s been through a typographical rollercoaster, you’re likely dealing with a forgery.

Syntax Slip Ups: Pay attention to the placement of colons, slashes and commas. Genuine cards have a standardized structure. Deviations might signal a counterfeit.

Logo Lapses: Examine the Aadhar and Government of India logos. Blurry or distorted emblems are telltale signs of a sham.

QR Code Quandary: A legitimate Aadhaar card includes a scannable QR code. Use the mAadhaar app or a QR scanner to verify if the embedded information matches the printed details.

If all else fails, you can always visit UIDAI’s official verification portal and enter the 12-digit Aadhar number. If it’s valid then you will receive confirmation.

Conclusion

India is only beginning to come to grips with innovative technologies such as AI and the images produced using it. While a large chunk of Indians on the internet are tech-savvy, elderly and senior citizens are particularly vulnerable to AI-created pictures, leading to misunderstandings and increased susceptibility to scams and misinformation.

Read: Instagram Starts New Experiments With AI-Generated Comments

Share: