Microblink Sets New Standard for Document Security with Perfect IDNet Test Results

BROOKLYN, N.Y.--(BUSINESS WIRE)--Microblink, a global leader in identity verification solutions, today announced the successful evaluation of its Microblink Platform solution against IDNet, the first large-scale, independently created dataset specifically designed for benchmarking document verification solutions against synthetic identities. The rigorous testing demonstrated the platform's strong capabilities in detecting AI-generated fraudulent documents, achieving zero false acceptances in the end-to-end analysis.

AI-powered identity verification through the Microblink Platform detects 100% of fake documents in DHS-backed IDNet dataset evaluation.

Share

Identity fraud is a growing threat, made worse by generative AI tools that make it easy to create convincing fake IDs with minimal effort. To combat this, the U.S. Department of Homeland Security sponsored the IDNet dataset, a robust benchmark for testing document verification systems against advanced synthetic fraud. It includes over 837,000 images from 10 U.S. states and 10 European countries, covering fraud techniques like face morphing, portrait swaps, and text field alterations.

Despite the increasing sophistication of these threats, there is currently no standardized metric for deepfake detection. Many identity verification vendors self-report their efficacy using different methods and metrics, leaving businesses unsure of how well their systems defend against Gen AI-fueled attacks. The IDNet dataset provides an objective standard for evaluating how well vendors perform. A trustworthy solution provider should be able to demonstrate real-world effectiveness by replicating a near-zero False Accept Rate when testing IDNet images.

"As generative AI makes it easier to create convincing fake identities, independent testing has never been more important," said Hartley Thompson, President and COO of Microblink. "Our results on the IDNet dataset show just how effective the Microblink Platform is at catching advanced synthetic fraud. We saw zero false acceptances in our comprehensive testing. This is a clear sign our AI is delivering real protection for businesses and consumers."

Microblink evaluated its platform for 10 U.S. states: Arizona, California, Nevada, North Carolina, Pennsylvania, South Dakota, Utah, Washington D.C., West Virginia, and Wisconsin, using 83,652 synthetically generated images within the IDNet dataset. To simulate real-world fraud more accurately, Microblink treated all samples, including those labeled as “bona fide,” as fraudulent.

The end-to-end analysis, incorporating all available fraud checks, showed zero false acceptances across all document types and states. This result highlights the Microblink Platform’s resilience against even the most realistic synthetic documents.

Some samples were flagged due to dataset-related image quality issues, such as over-cropping or missing fields, but none were incorrectly accepted as legitimate. A focused test on face photo tampering also produced strong results, confirming the system’s ability to detect synthetic faces beyond obvious visual cues.

“We’ve been using our in-house Fraud Lab for years to generate synthetic IDs and simulate real-world fraud,” said Thompson. “With over 280,000 AI-generated documents, we’re constantly training and testing our models to stay ahead of both digital and physical tampering. The release of the IDNet dataset gave us the perfect opportunity to validate our technology against a rigorous, third-party benchmark. The results speak for themselves.”

The full evaluation methodology is available in this report. Results will also be discussed in a free webinar.

About Microblink

Microblink builds proprietary AI solutions for identity verification and payment fraud, empowering businesses worldwide to improve customers’ digital experiences. With more than 12 billion documents processed from over 180 countries, the flagship Microblink Platform is used across industries to onboard more real customers, optimize KYC/AML workflows, and minimize fraud, enabling organizations to make online interactions safer and easier.