Insurers are turning generative AI against itself as more people use the technology to fake evidence for damage claims. Mitsui Sumitomo Insurance developed a system in August that judges whether photographs submitted as proof have been created by generative AI, and has begun using it, the Mainichi Shimbun reported.
Of 7,504 people who filed fire, injury and other insurance claims in September, the company judged 16 to have submitted images made by AI. The figure is on a rising trend from August, according to the report.
Fraudulent claims have generally been run by criminal organizations recruiting people through underworld job ads, a practice known in Japan as yami baito. Generative AI has made it easy to fabricate the images that serve as evidence of damage.
Insurers are struggling with what the Mainichi Shimbun described as a casual turn in fraud, in which people file false claims on a light impulse rather than as part of an organized scheme.