Selfies: The latest big data source for life insurance
As the old adage says, a picture is worth a thousand words. But for life insurers, the value may go beyond that.
That's the objective of a new insurtech startup, Lapetus, which has developed facial analytics software that uses selfies to analyze body mass index, gender, and even physiological age to determine underwriting risk.
The platform, Chronos, uses machine learning algorithms to extract human faces from images. Its AI engine then looks for intricate signals that give away information life carriers consider when issuing policies.
As an example, Lapetus deciphers BMI by looking at the roundness of the jaw and cheek area, while gender is determined by measuring the size of an individual’s eye orbit or mouth. The list goes on, according to Karl Ricanek, chief data scientist and co-founder of Lapetus. Chronos can also interpret smoking habits, drug use, and signs of depression.
“A picture is just a combination of ones and zeros,” Ricanek says. “When you walk into a doctor’s office, they are doing the same thing. They know your face is the window into your health.”
Lapetus is working to add the detection of genetic disorders, diabetes and heart disease to its AI software, Ricanek says. Several insurers are currently testing Lapetus’ technology, according to the company, though it denied sharing specific names citing non-disclosure agreements.
“We reduce months-long processes into minutes,” said Ricanek. “Now, insurers can price based on an individual’s risk and not larger risk pools.”
Lapetus employs 16 people, the majority of which are data scientists. Customers upload selfies to carriers, which are then sent to the Wilmington, North Carolina-based startup for testing using web-connected APIs. From there, its AI software spits out pertinent data insurers want. Additional questions from Lapetus or carriers could also be required for a more accurate risk profile.
“We have two buckets of tools. One is bio demography [used to determine life expectancy based on a series of questions about family and medical history]. The other is facial analytics,” said Chris Reeves, Lapetus’ chief technology officer. “Insurers determine both where the experience is hosted and what kind of package they want from us according to their underwriting needs.”
Lapetus has more than 1.5 million selfies in its database. A majority come from its time as Lapetus Software before re-launching in April 2015. The rest are made up by Ricanek’s 15 years of research and the startup’s ongoing human-subject-testing program, SMILe.
To ensure customers are comfortable with having their selfies analyzed, Lapetus also conducted two market studies, surveying a little less than 500 consumers. It found clients would have no issues with submitting images for life underwriting purposes; primarily because the experience feels more personal, it said.
“Certainly millennials are super comfortable with it,” said Reeves. “Once a company is convinced, it’s up to them to incorporate this.”