Ondato’s AI-enabled ID forensics system detects document alterations

Ondato launched a new AI-enabled ID forensics system that makes it difficult to submit spoof documents, such as altered photos, ID papers, and deep fake videos, as part of any customer on-boarding process.

Ondato AI-enabled ID forensics

The current scale of hacking and ID theft means that fraudsters have a better chance of finding a photo that matches their target scam. While these spoof documents may pass an “eye test”, Ondato’s AI algorithms notice any discrepancies in the data and flag them as attempted fraud.

By October 2021, nearly 281.5 million people had been affected by some sort of data breach. And in November, a hacker leaked a database for Argentina’s entire population – over 45 million individuals – divulging information that could be adapted to spoof less-advanced detection measures. The hacked data is now being sold in private circles.

Ondato’s new spoofing detection product protects legitimate customers by preventing stolen documents being used without their knowledge in identity theft hacks. Such risks are on the rise. The system automatically and instantly detects attempts to submit altered documents in an attempt to bypass identification measures and open a customer account fraudulently.

It provides banks, other financial institutions and any entity obliged to perform KYC processes with the highest available level of protection and the assurance they are on-boarding legitimate customers in full compliance with KYC and AML regulations.

Liudas Kanapienis, CEO and co-founder of Ondato, says: “ID theft and fraudulent account openings are a major and growing problem for individuals and for financial institutions too. If banks can’t trust customers’ ID documents, then they can’t be confident they are complying with local and international regulations. Ondato solves this challenge with new algorithms and machine learning that provide the most effective level of ID document checks currently available.”

Ondato is using algorithms and machine learning to detect document alterations. The solution detects various combinations of spoof attempts involving replays, reprints, image alterations and physical spoofing.

Ondato’s system checks if a video feed that claims to be the likeness of a new customer is in fact fake images being broadcast from a screen. This replay fraud is detectable in various ways, for example, if the screen has edges or there are flickering patterns. It protects you against live stream forgeries.

Attempts to alter documents are detected from irregularities between font sizes, the distance between letters, and artifacts – inconsistencies within an image’s compression as a result of attempted fraud. Image alterations are revealed by pixel distortion analysis that looks for pixels in the wrong places due to photo alteration, and alterations to uploaded files’ metadata.

Ondato also checks for physical spoofing, where alterations are made to part of a document, for example, if another image or part of an image has been glued on or physically printed on a piece of paper. Techniques here include analysing the image depth, the registry data and unexpected shadows.




Share this