Finatext has launched a proof of concept (PoC) with Medicare Life to detect insurance claim fraud by analyzing signs of generative AI manipulation remaining in claim images. The images covered include medical certificates, receipts and medical statements; details of the analytical methods have not been disclosed.
Finatext Holdings has launched a proof of concept (PoC) with Medicare Life Insurance, part of the Sumitomo Life Group, to use AI in detecting insurance claim fraud. The PoC analyzes signs of generative AI manipulation remaining in claim images, with the aim of identifying fraudulent claims at an early stage.
The initiative comes against a backdrop of increasingly sophisticated methods for altering and falsifying claim documents as generative AI and image-editing technologies advance. Balancing the rapid payment of legitimate claims with the early identification of fraudulent claims has become a challenge. Medicare Life and Finatext have previously advanced the use of AI in claims assessment and achieved high accuracy in document reading, including through AI-OCR.
The PoC covers claim images such as medical certificates, receipts and medical statements. AI will be used to detect signs of generative AI and assess the possibility that documents have been altered or falsified. The analytical methods have not been disclosed. The aim is not for AI to make the final determination of whether a claim is fraudulent, but to improve the efficiency of claims assessors' review work.
Going forward, the companies will use the PoC to validate the effectiveness of detecting signs of manipulation, consider integrating the technology into claims assessment operations, and organize issues related to data preparation, with the aim of enhancing insurance operations and strengthening the ability to respond to fraudulent claims.