Viewpoint: Detecting Digital Deception in Claim Photos

More than 85% of Americans own a smartphone, meaning about 300 million people across the country have high-resolution digital cameras tucked into their pockets. The ubiquity of these devices has led to efficiencies in the claim-handling process for many property and casualty carriers. By enabling insureds, claimants, providers, and others to capture and share photographic evidence in support of claims, insurers can investigate and process those claims with increased speed. One study indicated that 27 percent of repairable vehicle claim estimates are initiated via photos.

These estimates seem like a win-win for insurers and customers. Policyholders can get faster, simpler service when reporting losses, and carriers reduce expenses without the need for onsite estimates. But there’s another group benefiting from remote claim estimates—fraudsters.

Digital photo technology has armed fraudsters with new tools and methods to defraud insurers in the claim-handling process. As carriers embrace customer and claimant-submitted photos, including photo-based estimates, and accelerate toward straight-through processing, they must reckon with the growing digital photo fraud problem.

The growing photo fraud threat

How big of a problem is insurance fraud? According to the Coalition Against Insurance Fraud, it costs the U.S. economy $308.6 billion each year. That’s nearly $1,000 for every person in the country.

Fraud is not a new problem for the industry, but in light of current economic headwinds and a nearly $27 billion net underwriting loss in 2022, reducing the leakage is critical. Certain factors have contributed to fraud’s growth, including increased digitalization, customer-focused claims strategies, and lackadaisical consumer attitudes toward insurance fraud.

The Internet and digital claims processes have made it easier than ever to commit fraud, and there are numerous ways to submit fraudulent claims using loss photos. Many of them are relatively simple. For example, fraudsters can submit the same image to different carriers as evidence of losses. A fraudster can also download a picture of damages from the Internet and submit it as evidence of their claim.

A classic example occurred in a McDonald’s coffee burn claim. In that case, a California woman submitted hand burn photos she had downloaded from a Colorado hospital website. Unless a hyper-vigilant adjuster happens to discover the duplicate image, these schemes will likely go undetected.

Fraudsters are also turning to technology to perpetrate schemes. Using photo-editing software, they can manipulate documents and images to fabricate or inflate damages or even create fake claims altogether. And with the emergence of generative artificial intelligence and certain smartphone apps, anyone can create or doctor loss images.

It’s not just claimants and insureds that perpetrate these schemes. Some third-party suppliers such as independent adjusters, repair contractors, and auto body shops might re-use images or download photos from the Internet to inflate bills or cut corners on providing estimates. For example, in a recent analysis of a carrier’s claims, we discovered that a single property appraiser used the same photo in 170 different claims over two years. This one appraiser’s nefarious actions impacted claims totaling more than $1 million in indemnity payments.

Insurers can protect against this growing source of leakage with help from digital forensics. This defensive technology uses AI to identify indicators of fraud and manipulation in digital photos and documents. A robust digital forensics program monitors and checks four primary threat vectors: 1) metadata (or Exif data) inconsistencies, 2) image duplication, 3) Internet sourcing, and 4) pixel manipulation.

Metadata and Exif data forensics identify inconsistencies in information embedded within an image file. Embedded information may include the date, time, and location the photo was taken, the manufacturer and camera model used to take the photo, the date a photo was edited, and other details. An example of an inconsistency in metadata is a substantial distance between the location a photo was taken and the loss location.

Image duplication occurs when someone submits the same image as evidence in two or more unrelated claims. This can occur by submitting the same photo to different carriers or to the same insurer under multiple policies and claims. For duplication detection to be effective, image forensics must compare photos across claims and carriers.

Sourcing photos from the Internet is often the easiest for a fraudster. Effective image forensics must not only search for duplication across P&C claims but also must identify images duplicated from the web.

Finally, advanced image forensics look beyond metadata and duplication flags. Forensics designed to detect pixel manipulation, for example, compare pixel patterns and markers inherent in the original photo to those pixels added or changed by software. Although not visible to the human eye, photo-editing software leaves trace evidence in the form of noise patterns easily spotted by well-trained AI forensic software. These forensic technologies provide insurers with vital capabilities in their anti-fraud toolkit.

As digital images become more and more central to the claim-handling process—especially straight-through-processing—it’s essential for insurers to employ digital forensics technology. Savvy fraudsters and opportunists have many ways to game the system with digital images, so insurers must be equipped with the latest techniques to quickly identify and scrutinize non-meritorious claims.

But the ultimate benefit of image forensic technology is not just for detecting fraud, it’s for processing legitimate claims more quickly. Carriers equipped with effective tools to verify the authenticity of loss photos can confidently and accurately pay meritorious claims faster, thus shortening claim cycle time and improving customer experience. That’s a true “win-win” for the carrier and policyholder.