Straight-through processing is considered the holy grail of digital claims—where insurers can intake, process, and pay claims automatically, delivering the rapid digital experience consumers expect today. Many carriers are progressing toward this goal in hopes of fully automating nearly half of all non-injury claims by 2025.
But there are risks to creating that degree of speed and digital enablement, mainly digital fraud.
We’ve been examining the top schemes in digital image fraud and how binary and non-binary forensics can help detect them. In the final blog post in this series, we’ll explore more sophisticated fraud tactics and how advanced forensics can identify them.
One of the ways fraudsters are scamming insurers is by manipulating documents and photos submitted with claim files. Documents, in particular, are easy to manipulate and don’t require advanced technology or special skills to modify.
For example, if someone wanted to submit a fraudulent invoice for reimbursement for a property claim, they could use free online PDF editing software to modify an original document. They could simply change the price on the receipt to inflate the claim.
They could even pass the same invoice to a friend or family member who could change the name and contact details on the invoice and use the same document to defraud another insurer.
This scheme can easily go undetected because the naked eye can’t identify modifications. Plus, there’s no way to examine each document at scale with the volume of claims adjusters handle.
Manipulating loss images
Manipulating image pixels is another scheme that is becoming easier to perpetrate with popular photo-editing software or even free online tools. Fraudsters can modify a few pixels of a photo to show damage that doesn’t exist.
The photos below show an example of image splicing, in which a region of one photo is copied and pasted on top of another. The image on the left is manipulated, as the damage was spliced on the original photo (the image on the right). It’s nearly impossible for humans to recognize these types of modifications with the naked eye.
Identifying fraud with advanced forensics
As fraudsters become more advanced, there’s a need for more sophisticated anti-fraud technology. Advanced forensics uses AI to identify manipulation in digital media files such as PDF documents and photos.
This technology goes beyond what the human eye can detect to uncover modifications. For altered documents, it detects where the manipulation occurred and reverts to the original document. So, if the price of a receipt was altered, it detects that the price changed and can often show you the price on the original document.
To detect image manipulation, the technology identifies noise patterns in an image. When cameras capture images, they compress them into a format, such as JPEG, creating noise patterns. Each camera has slight nuances in the noise patterns. At Verisk, we’ve developed technology that identifies if an image has more than one noise pattern, which is a sign of manipulation. The images below show how the technology identified a different noise pattern in the manipulated bathroom damage photo.
Innovation to streamline claims
Digital media forensics is an essential element in insurers’ journey to straight-through processing. Without this type of technology, companies leave themselves open to significant fraud exposure.
The ultimate benefit of forensic technology—whether it’s binary forensics, non-binary forensics, or advanced forensics—is for processing legitimate claims. These tools allow carriers to automatically examine loss photos and validate the date and location the image was captured, determine whether the image is from a prior loss, and confirm that documents and photos are original. Those tools, along with a ClaimSearch® loss history review, give carriers the confidence to pay meritorious claims quickly, which boosts customer satisfaction, shortens cycle times, and reduces costs.
To learn more about Verisk’s digital forensics suite, contact Jim.Hulett@verisk.com.