A Blog by Jonathan Low

 

Jun 14, 2019

How An Insurer Is Using Facial Recognition Technology To Assess Creditworthiness

The eyes are the window to the soul. Any maybe to the pocket book. JL

Zhou Wei reports in the Wall Street Journal:

The company has used it to screen individuals applying for loans. “Micro expressions,” brief, small, involuntary movements on people’s faces, reveal their emotional and psychological state. Such expressions are hard for people to control. Loan officers make more accurate judgments on applicants’ credibility based on this information. For large loans, applicants answer questions in an online video. (It) analyzes how the applicant answers, and looks for signs of eye-shifting or suspicious behavior. It has made 500 billion yuan worth of loans with micro-expression technology, which has helped shorten loan-approval times to two hours from five days.
What does your face reveal about how big of a financial risk you are?
One of the world’s largest insurers thinks the answer is quite a bit. It is using facial-recognition technology on potential customers as part of its efforts to assess risk when it sells them financial products
Ping An Insurance Group Co., China’s largest insurer and one of the country’s largest financial conglomerates, routinely scans the faces of customers and its own agents to verify their identities or examine their expressions for clues about their truthfulness.
Since 2016, the company has used its proprietary technology to screen individuals applying for loans at its consumer-finance arm. “Micro expressions,” or brief, small and often involuntary movements on people’s faces, are one area of focus the company has been analyzing.
The efforts are a reflection of what’s going on throughout China, where facial-recognition technology has become a feature of daily life. Unfettered by regulations or privacy concerns, Chinese authorities use it on public streets, subway stations, at airports and at border controls.Chinese insurance companies may face fewer privacy concerns from consumers, compared with their Western counterparts, says Tjun Tang, a Boston Consulting Group senior partner in Hong Kong, though he adds that Chinese consumers are becoming more concerned about privacy in recent years. “If the consumers are willing to share the data, that’s helping a lot,” he says. “It allows machines to learn.”
Ping An uses its facial-recognition technology mainly to verify customers’ identities. When people open accounts with the insurer, they are often asked to submit images of their identity cards and complete a series of tasks including opening their mouths and blinking when they use the company’s facial-recognition program for the first time.
After setting up the accounts, customers can purchase some life- and health-insurance products after scanning their faces with their smartphones. They can also connect directly with insurance agents in video meetings or submit claims after Ping An’s software recognizes their faces.
Shi Haojing, a 28-year-old airport worker in Shanghai, says she scans her face with her smartphone camera when she needs to speak with an insurance agent and collect or make payments on her Ping An policies.
“It is good that things can be done without leaving home,” she says. “Everyone is getting lazier.”
Ping An also uses the technology in its insurance business to gauge customers’ health. In one new application, facial scans are used to estimate the body-mass index, or BMI, of customers who want to purchase a policy that makes lump-sum payouts of up to one million yuan (about $145,000) when they are diagnosed with any of 100 critical diseases. Policyholders get discounts on their monthly premiums based on how much body fat they have, as calculated by the scan. Individuals judged to have a BMI under 30 can receive a monthly discount of 6 yuan (about $0.87) on the product, which the company sells at a starting price of 30 yuan a month. A BMI of 30 or higher is considered obese.
“Just the face itself contains a wealth of information about an individual’s health,” says Michael Powers, a professor of risk mathematics at Tsinghua University in Beijing. It may be possible to glean from a person’s face whether the individual is a smoker, for example, he says. But such use of facial-recognition technology hasn’t yet been commercialized widely, in part because there are concerns it could be used to discriminate against some groups of people.
Some critics say the technology isn’t really being used to set prices for different groups but instead is a kind of marketing tool for the company to generate attention and make consumers more familiar with AI.
Cliff Sheng, a Hong Kong-based partner at consulting firm Oliver Wyman who has closely studied uses of technology by Chinese insurance companies, argues that there are more reliable ways than facial recognition to acquire users’ BMI. However, Mr. Sheng does say he thinks that the technology can be useful to insurers as a marketing tool because it gets people’s attention.
A spokesperson for Ping An says that its BMI-measuring feature using facial recognition is provided to the consumers as “a benefit,” and that its underwriting process takes into account a variety of other factors and data points.
In its lending business, meanwhile, Ping An says it uses its technology to analyze the faces of loan applicants in real time, searching for “micro-expressions” that reveal their emotional and psychological state. Such expressions typically occur within fractions of seconds and are hard for people to control, and loan officers make more accurate judgments on the applicants’ credibility based on this information, according to an article posted by Ping An on its official WeChat social-media account in China last year.
For large loans, applicants often have to answer questions in an online video meeting that typically lasts 10 to 15 minutes. Ping An records and analyzes how the applicant answers questions, and looks for signs of eye-shifting or other suspicious behavior, which would be flagged by its system.
Ping An in January said it has made more than 500 billion yuan worth of loans with the help of its micro-expression technology. It also said the technology has helped shorten its average loan-approval times to two hours from five days.

0 comments:

Post a Comment