Skip to content
Submit incident
AII-1001-87233
ReportedReported by fewer than three independent publishers, with no official finding.Supporting sources are pending verification.

ChatGPT-4o's image generator enables realistic fake receipt creation

ChatGPT-4o's improved image generator can create realistic fake receipts with sufficient detail to potentially be used for expense fraud. OpenAI says generated images include metadata identifying them as AI-created and notes it enforces its policies against fraudulent documents, but has not explained why the capability is permitted.

31 March 2025 Β· OpenAI Β· Fraud, scams and manipulation
31 March 2025
AII-1001-87233
Quick facts
Organization
OpenAI (subject)
Category
Fraud, scams and manipulation
Country
United States
Date
31 March 2025
Drafted by
Raidu Incident Detection

What happened

ChatGPT's 4o model, unveiled in March 2025, includes an improved image generator that can create realistic text within generated images. Users have demonstrated the ability to generate fake receipts for real restaurants, with some examples including visual details like food stains and crinkled paper to increase authenticity.

Multiple users on social media posted examples of AI-generated fake receipts, including one for a San Francisco steakhouse and another for an Applebee's location. TechCrunch tested the generator and confirmed it could produce fake receipts, though with occasional errors such as incorrect math and formatting inconsistencies.

Reported impact

Security researchers and tech commentators have expressed concern that the technology could be misused for expense fraud, enabling bad actors to submit reimbursement claims for entirely fabricated expenses. The ease of generating convincing fake receipts is described as presenting 'huge opportunities for fraud.'

Index analysis

OpenAI spokesperson Taya Christianson stated that all images generated by ChatGPT include metadata indicating their AI origin, and that OpenAI 'takes action' when users violate usage policies. OpenAI's policy prohibits generating fraudulent documents, though the company has not explained why the image generator permits receipt generation in the first place.

Classification

Organization
OpenAI (subject)
Category
Fraud, scams and manipulation
Industry
Financial and insurance services
Country
United States
Harm status
Potential harm
Intent
Unintentional
Timing
Post-deployment
Harm types
Economic or property
Harmed
Business
AI systems
ChatGPT-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