Plus, subtle AI edits to claims photos
A fraud probe sets its sights on a fake insurance agent license testing center down in Miami-Dade, and subtle photograph tweaks using artificial intelligence are causing an uptick in claims fraud. It’s all in this week’s Fraud News.
Fraudulent Insurance Testing: Last week, investigators arrested 12 people in Miami-Dade County for obtaining illegal insurance licenses through D and R Financial Services, a testing center that has been under extensive investigation since its owner’s arrest back in 2024. Law enforcement initially booked Rainier Salas for operating a criminal enterprise that held fraudulent insurance testing and licensing courses – now the whole house of cards has come tumbling down. More than two years later, investigators have tracked down 12 individuals who were illegally licensed through the facility, where leadership was accepting payments to take exams on behalf of the applicants, undermining the merit of Florida’s insurance agent licensure system. You can find the list of those arrested in this NBC 6 South Florida story. All have been charged with organized scheme to defraud and a list of other charges, and currently await their date in court.
“Vanilla” AI Edits: Most people worry about “deepfakes” and other high publicity uses of AI, but adjusters and investigators in the insurance space are worried about the impacts of “vanilla synthetics,” where slight edits to photos are costing insurance companies big time. New analysis from SAS is showing that small edits and fabrications from generative AI are getting more and more convincing, manipulations which run up claim valuations, passing many human reviews within the system. While technology is currently being employed to combat the fraudulent use of AI, fraud experts say the issue is rampant and will only get worse. Look no further than a challenge posited in a recent article from Insurance Business Magazine: two of the images are AI fakes, and one shows real damage. SAS said 40% of people couldn’t spot the fakes. There are some telltale signs of doctored images, like strange shadows, damage inconsistent with impact, and strange blurriness to an image, but machine learning is getting better at generating images by the day. In a world where fraudulent claims have added around $67 to the average premium, and one in seven claims are fraudulent, the numbers speak for themselves. In the meantime, the insurance industry has to hope that the machine learning they employ can offset the bad guys.
LMA Newsletter of 3-9-26
