TruthScan today announced an update to its AI image detection model, which helps retailers and online marketplaces identify AI-generated and AI-edited damage photos submitted with refund claims.
The update adds training on additional image generation sources, supporting fraud and claims teams reviewing customer-submitted evidence.
- The updated model can spot AI-generated images even when C2PA data or SynthID watermarks are removed.
- Testing puts TruthScan’s detection accuracy at 99.3%.
- In an analysis of 27,440 images involving damaged items or products submitted to TruthScan, it flagged 32.8% as being AI-generated or AI-edited.
“For marketplaces like Depop, Vinted and Mercari, trust between buyers and sellers is foundational. Our goal is to help fraud teams distinguish faked evidence from legitimate claims before a refund decision is made–at scale,” said Christian Perry, CEO of TruthScan.
TruthScan combines metadata and watermark analysis, pixel-level analysis, and a proprietary machine learning model trained on millions of images.
TruthScan’s core image detector is available for trial and standard subscription licenses. It also offers custom enterprise setups tailored specifically to their needs. To inquire, submit a contact form here.
About
TruthScan provides enterprise AI detection and fraud prevention solutions for images, video, text, voice and documents. It serves fortune 500 companies and has customers from finance, e-commerce, government and more.
Media Contact
Devan Leos
Chief Communications Officer
