Can Artificial Intelligence (AI) Decide Insurance Claims?
Posted in Lawsuit on November 30, 2017
With the recent increase in fraudulent insurance claims, coupled with the already present workload that authentic claims entail, insurance companies have explored several ways to help ease the load on their workforce, improve speed, and cut down costs. Enter technology and the automation of tasks computers offer. One company in Switzerland, however, is taking this step even farther, employing a form of artificial intelligence to the insurance claims process in hopes of cutting down the costs and time exponentially.
The machine takes advantage of machine learning to improve and upgrade its accuracy and operating efficiency. The process involves the machine scanning through personal injury claims, medical records, and other pertinent documentation. The program considers the evidence in relation to the claim’s basis to come to a decision. With each case, the machine learns, tracking trends to help it predict what is coming.
The intelligence would have the means to spot a growing trend in claims that could be a spike in fraudulent claims or the advent of a change in the marketplace, which would give the insurance company a head start in adapting policies to cover these new claims.
Lack of Emotion
The most obvious advantage to having a machine do this kind of work lies in the removal of the human element. Claim workers are subject to tiredness, weariness, disdain, or distraction. They require care that a machine does not. A claim worker’s perspective could be tainted or twisted by a previous experience with such a claim or by their own experiences with an event leading to a claim. A computer would not experience this.
The insurance company in Zurich, Switzerland employing this technology already sees benefits in using the artificial intelligence. One estimate they report is that the machine saved the company over 40,000 working hours, thus saving the company hundreds of thousands, if not millions, in expenses. In a market with an almost non-existent margin of profit, this kind of savings would prove very appealing. The company also reported cutting the process down to as little as five seconds, a strong selling point to those filing claims that typically wait much longer for decisions.
The idea that the machine learns by using previous information it encountered leads many to suspect the company retained the confidential information. This creates a target for potential hacks to steal that data, but also raises red flags on privacy concerns. Critics point out that this kind of operation would need strong external oversight to ensure proper protection of customer information and proper use of the data in ways that do not violate customer rights. If the idea of a human claims worker shuffling through previous clients’ information to inform themselves on a current claim raises concerns, the same should be true of a computer.
Employing technology to improve processes stands to benefit both the insurance companies and customers. Improved operating speeds and accuracy lead to correct decisions for the cases that are valid, while eliminating the number of fraudulent claims that take up company hours and expenses better spent on valid claims. However, these improvements do not absolve the company from addressing the valid concerns over what they do with the data. As it stands now, the potential for improvement exists, but only time will tell if these advanced computers can really deliver on the promised results.