AI Might Be The Solution To Insurance’s Costliest Problem
Machine learning (ML) and artificial intelligence (AI) systems have been developing at breakneck speed over the past several years. They’ve been turning up in industries far and wide, helping to solve some of the longstanding problems that were once thought to be insurmountable. The latest industry to feel the effects of the growing ML and AI wave is the insurance industry, and there are some growing signs that the latest technology could be the solution to the industry’s biggest – and costliest – issue.
A true cost driver
For individuals and businesses alike, insurance is a necessary bulwark against financial losses due to property theft, damage, and loss. It’s also a critical part of the healthcare infrastructure of most industrialized nations. Insurance, in general, is one of the few universal products that reach into every corner of our personal lives as well as throughout every economy around the world. That means that the cost of maintaining coverage has a direct effect on economic outcomes and even quality of life for those paying for coverage.
It may come as a shock that up to a third of insurers report that fraud accounts for as much as 20% of claims costs. In fact, the FBI reports that fraud costs the US insurance industry more than $ 40 billion per year, and that figure excludes the health insurance sector, where the incidence rate is believed to be much higher. That translates into hundreds of dollars in additional premium costs on every policy issued, and it’s a significant drain on the global economy, as well as on the wallets of individuals everywhere.
AI joins the anti-fraud fight
To combat the longstanding pattern of fraud and abuse in the insurance industry, a new breed of insurance technology firms are turning to AI for insurance claims processing. One of the early purpose-built AI systems, known as Force, is already in wide use by insurers throughout Asia and Europe and has already shown a success rate that is 2.5 times the current industry average. The key to the system, of course, is the variety of data it analyzes to spot potential fraud, meaning that the success rate should continue to climb as the system examines more and more claims.
Direct claims examinations aren’t the only way that AI is being used to root out potential insurance fraud, though. Companies like Hanzo are also working with insurers to deploy machine learning systems that make use of data streams from things like social media posts and online marketplace sites to look for signs of fraud. For example, the systems can look for listings on eBay that match items that had been reported stolen or find activity posted to social media that suggests that a claimant might be fabricating an injury or illness. It’s the kind of deep-dive into the Internet that would be impossible for a human claims investigator to accomplish with any efficiency.
The start of a revolution
If ML and AI systems manage to turn the tide against fraudulent claims in the insurance industry, it would have a significance that reaches well beyond the insurers themselves. The accrued savings across all insurance types would likely add up to a sum that would dwarf the 2009 stimulus package that helped halt the last recession in the United States. It would also bring an end to one of the most economically damaging problems that any global industry now faces, further cementing ML and AI as the technology most likely to shape the 21st century and beyond – and that’s an outcome that everyone should root for.
To learn more about industries that are making use of the latest in AI, read From Farm To Table: Bringing AI Into An Enterprise.