(Yicai Global) Sept. 21 -- Big Data can detect those who lie about their incomes when buying wealth management products, industry professionals said. This is thanks to colossal customer databases held by big internet or insurance companies, they added.
"About 65 percent of users lie about their incomes when buying wealth management products, but such dishonesty can be detected by Big Data," said Yang Jun, chief risk officer of Ping An Insurance (Group) Company Of China [SHA:601318], during a meeting with the representatives of the media.
Ping An Insurance controls data about 880 million Chinese users, disclosed a senior Ping An executive at a scientific and technological achievement conference yesterday. Data collected via biological recognition, Big Data risk control, blockchain, artificial intelligence (AI) and cloud technology can help financial institutions identify qualified investors.
Ping An Insurance has about 540 million financial customers, plus 400 million users in its Internet sector initiatives like Autohome. Some 300 institutions, including UnionPay, also cooperate with Ping An Insurance. Currently, the user data enables the Group to determine customers' credit lines from static aspects such as age, income, and gender, as well as dynamic aspects like increases in mobile expenditure, said Ou Haiying, chief product officer of Ping An Technology (Shenzhen) Co., a wholly-owned subsidiary of Ping An Insurance (Group).
Ping An Easy Money, a financial service cluster affiliated to Ping An Insurance, is a case in point. At present, loans from Ping An Easy Money can be provided within 3 minutes at its fastest speed. In this process, the system first detects whether the applicant is a real user based on identity recognition, and then the system determines whether to authorize the loan and how much to provide, based on a credit information report. Finally, the customer secures the release of the credit with an electronic signature, Ou said.
Recently, Ping An Insurance has focused on intelligent cognition field in its scientific research, Ou stated. As there have been cases where biological recognition based on a single feature such as fingerprint and face is cracked, Ping An Insurance is, therefore, trying to combine vocal print recognition with facial recognition and identity recognition in order to enhance the reliability of identification in financial scenarios. In the future, Ping An Insurance may also research identity recognition based on veins and irises, Ou added.
At present, facial recognition only addresses the question of "who you are." The next technological step will be to analyze humans' micro-facial expressions. Ping An Insurance has obtained 19 micro facial expressions so far, Ou said, which can be used in combination to help identify seven different emotions. For example, there can be a happiness index or a nervousness or sadness index, and these indices can be used to help make business decisions. For example, when a customer feels very nervous when answering questions from a financial institution, the institution will ask further questions to find out more details from the customer.