US EditionPublished • loading... • Updated
AI-Generated Fake Receipts Cost Companies Millions
Nearly 70% of CFOs suspect employees use AI to falsify expenses, and AI-generated fraud cases are expected to rise, experts say.
- In recent months, business expense software providers report a surge in fake receipts turned in by employees using AI to fabricate expense claims.
- In recent months, major AI companies including OpenAI and Google released new image-generation models, while OpenAI's GPT-4o image update in March triggered a sharp rise in AI-made receipts.
- AI programs can produce receipts with realistic imperfections like wrinkles, smudges, and signatures, while detection software inspects hidden metadata and pixel patterns showing some images were made by ChatGPT.
- AppZen said about 14% of fraudulent documents were AI-generated last month, Ramp flagged over $1,000,000 in fraudulent invoices within 90 days, while large expense tracking platforms process more than 80 million compliance checks each month.
- In response, businesses are tightening expense policies and exploring blockchain verification while preparing for metadata-scanning tools, though metadata can be removed, so no clear solution exists, reports AppZen and Ramp.
Insights by Ground AI
Podcasts & Opinions
30 Articles
30 Articles
Workers using AI to scam their employers with phony expense receipts: report
Artificial intelligence is helping workers create fake expense receipts, according to a recent report. A slew of AI-generated receipts have been submitted to employers thanks to new image generation models from companies like Google and OpenAI, the Financial Times reported.FOX Business reached out to Google and OpenAI for comment.Software provider AppZen told the outlet, for instance, that these fake AI receipts accounted for roughly 14% of all …
·United States
Read Full Article
Coverage Details
Total News Sources30
Leaning Left2Leaning Right1Center6Last UpdatedBias Distribution67% Center
Bias Distribution
- 67% of the sources are Center
67% Center
L 22%
C 67%
11%
Factuality
To view factuality data please Upgrade to Premium
