Had a discussion with my brother in law yesterday. He's a private investigator who specializes in insurance (auto/bodily injury)fraud. He gets potential fraudulent claim referrals from the insurance companies which he then investigates. His tales of the lengths that people go to defraud insurance companies are legend, including policy holders that injure themselves or a family member to make a bodily injury claim.

He recounted that recently the insurance company he works for (a name you would recognize) had begun to use artificial intelligence to make the fraud referrals to him. The introduction of artificial intelligence has had resulted in a much higher "hit"(success)rate. AI improves the use of his time by bringing forward more easily to prove fraud cases. He claims that, in general, he is spending half as much of his time on each case. AI is amazingly fast and efficient scouring public records, social media, proprietary databases, etc.. looking for anomalies and inconsistencies.

AI can compile the data and estimate the probability of fraud using a number of different factors. Brother in law then uses the AI to prepare a "likelihood of fraudulent claim report" (paraphrasing) that is then sent to policy holder and/or their attorneys. Once they receive the AI generated report more of the claimants are now deciding not to pursue their claims.

He's concerned that the AI will probably replace his position entirely. I suspect he is right.