Skip to content
Submit incident
AII-1001-87873
ReportedReported by fewer than three independent publishers, with no official finding.Supporting sources are pending verification.

ChatGPT GPT-4o generates realistic fake Aadhaar and PAN cards

Within a week of OpenAI enabling image generation in GPT-4o, users generated realistic fake Aadhaar and PAN cards, raising concerns about potential misuse for fraud. While the fake documents contain detectable differences from authentic ones, they reportedly appear genuine enough to pose security risks.

4 April 2025 Β· OpenAI Β· Fraud, scams and manipulation
4 April 2025
AII-1001-87873
Quick facts
Category
Fraud, scams and manipulation
Country
India
Date
4 April 2025
Drafted by
Raidu Incident Detection

What happened

A week after OpenAI unlocked the image generation feature in GPT-4o, users began generating fake Aadhaar and PAN cards with striking similarity to authentic documents. Over 70 crore images have been generated using the feature since its launch.

Social media users shared images of fake Aadhaar cards bearing the names and faces of notable figures such as Sam Altman and Elon Musk. The fake documents included QR codes and Aadhaar numbers that appeared genuine, though the QR codes could not be scanned to retrieve holder information.

Reported impact

The generated fake identity documents have triggered concerns about potential misuse for fraud and identity-related crimes, as Aadhaar is a sensitive ID issued by India's UIDAI. OpenAI acknowledged that the image generation capabilities introduce new risks distinct from previous generative models.

Index analysis

While the fake Aadhaar and PAN cards are visually similar to real documents, there are detectable differences including variations in font, punctuation, syntax, placement of text elements, and distortion of logos such as the National Emblem. The QR codes in fake documents cannot be scanned to access information from the UIDAI website.

Classification

Category
Fraud, scams and manipulation
Industry
Government, security and defence
Country
India
Harm status
Potential harm
Intent
Unintentional
Timing
Post-deployment
Harm types
Economic or property, Public interest
Harmed
General public, Government
Failure modes
Deepfake image
AI systems
GPT-4o

Sources and evidence

Sources not yet published

This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.

How sources are verifiedSuggest a source for this incident