人工智慧伎倆與虛假索賠：保險詐欺的新面孔

Insurance companies in Wales are facing a new problem: people are using artificial intelligence to create fake damage and imaginary items in their insurance claims. Admiral, a big insurance company based in Cardiff, has seen a 71 percent increase in fraud in 2025 compared to the year before. They say that more people are using AI software to change photos and make up documents to try to get money from insurance. For example, some people have sent in pictures of expensive watches that were actually created by a computer, not real watches. Others have changed the number plates on cars in photos to try to make the same accident look like two different claims.

The insurance industry is very worried about this new kind of cheating. Haith, who works in Admiral's household claims team, explained that people are using AI to make it look like their car or house was damaged in a certain way, or to create documents that never existed. If someone is caught making up or exaggerating a claim, they could have their insurance cancelled, their claim rejected, or even face criminal charges. Admiral and other insurance companies are now investing in new technology to help them spot fake images and documents. They have special software that can tell if a photo has been changed by a computer.

John Davies from the Insurance Fraud Bureau said that not only are regular people using AI to exaggerate their claims, but organized crime groups are also using it to make fake documents and cheat insurance companies. The industry is working together to fight this problem by sharing information and best practices. Flora, who works on Admiral's fraud team, said that the consequences for getting caught can be very serious. People who cheat could end up with a criminal record, higher insurance costs, or other problems that could affect their lives for a long time. She warned that trying to cheat the system is simply not worth the risk. As insurance companies get better at detecting AI-generated fraud, they hope to stop more people from trying to scam them.