Mitsui Sumitomo Insurance and Aioi Nissay Dowa Insurance have introduced an AI-powered image detection tool designed to help identify potentially manipulated claims documents, as insurers step up efforts to address the growing risk of fraud involving generative AI.
The two insurers, part of the MS&AD Insurance Group, jointly developed the technology with Tech Union.
According to the companies, the system analyses multiple electronic characteristics and metadata embedded in images submitted as supporting documents for insurance claims, including photographs of damage, repair estimates and receipts.
The detection function is designed to identify indicators that an image may have been generated or processed using AI, as well as signs of image manipulation or the reuse of previously submitted images. The results will be used as reference information by claims handlers during the claims assessment process, with final payment decisions continuing to be made by claims personnel.
The insurers said they are the first domestic insurers in Japan to introduce such a capability into insurance claims operations.
The initiative comes as advances in generative AI make it increasingly easy to create highly realistic images, raising concerns that the technology could be exploited to support fraudulent insurance claims.
At the same time, insurers are receiving a growing proportion of claims documentation through digital channels, increasing the need for tools capable of detecting manipulated or fabricated evidence.
The companies said the introduction of the detection function will not change existing claims procedures or create additional requirements for policyholders.
The technology, which is currently the subject of a patent application, represents the first phase of the group's broader strategy to strengthen its defences against AI-enabled fraud. The insurers said they plan to further enhance the detection capabilities as AI technology continues to evolve.