Face-powered payments are coming

Facial recognition has been around for a few years, and its primary use is in situations where your face is compared to only one image, the one stored in your ePassport or your digital ID. Using facial recognition to identify a face among thousands or millions of others is a way more complex story!

Around the world, an increasing number of retailers are toying with the idea of making consumers pay directly with their face.

In Russia, multiple supermarkets and cafes in Moscow are proposing a “pay by face” option to their customers. The X5 Retail Group, a major food retailer has partnered with Visa and Sberbank, a major Russian financial institution, to implement a 'pay-with-a-glance' option at around 150 of their Perekrestok supermarkets. Consumers simply enroll their face on their Sberbank app on their smartphone and link it with a given account or credit card. Then, they present their face to a camera when exiting a store, are recognized and directly debited from the right account. In an interview to Finextra, Kirill Tsaryov, Sberbank's head of retail business, says: “Our research confirms that this type of biometric identification is gaining traction in Russia, with 70% of customers planning to rely on face payments going forward.” Biometric payment technology may be expanded to some 3,000 X5 stores by the end of 2021. The solution developed by X5's innovation lab technology and Netherlands-based VisionLabs solutions, uses a 3D camera with high detection accuracy and depth capture, to ensure an easy and quick facial recognition.

This innovation is far from being limited to Russia! For instance, in California, Piestro, a company that sells robot-made pizzas, goes to the extent of automatizing payments thanks to a face recognition technology developed by PopID, a Los Angeles based startup. With this technology, consumers first have to enroll their face in the app, and then, when they are in a given retail location, the stand in front of a camera, get recognized, have their face associated with their order, and complete their payment automatically. PopID technology is deployed at a few restaurants in California, using simple smartphones or tablets as cameras. In addition to its facial recognition technology, PopID acts as a wallet operator: they charge the consumers first, own the corresponding float and pay merchants with a processing fee lower than traditional card transactions, according to Kiosk Marketplace.

The “pay by face” paradigm is attracting many newcomers to the payment industry. PayByFace, a Romanian start-up has deployed its biometric facial authentication-based payment system, validated by the National Bank of Romania, in coffee shops and a few pharmacies in Bucharest. In Russia, SWiP, a Russia-based startup has developed a full “SelfieToPay technology,” which has been tested with Raiffeisenbank. Also, Mati, a start-up created in San Francisco to develop facial recognition technologies, has moved to Mexico to be closer to dynamic Latin American markets. They started as a service for identity verification online built to help businesses comply with KYC / AML regulations, and are now proposing a full system for face-based biometric payments.

Pay by face technology can only be implemented on a wider scale when it is demonstrated all safeguarding procedures for privacy have been implemented. For the time being, data protection activists have already expressed doubts the enrollment for payment in stores or on banking apps could be used for other purposes such as identifying people on the street in authoritarian regimes.

For the time being, these “pay by face” systems still have to overcome potential regulatory obstacles and be adopted by the public at large. Like many innovation in payments, they no longer rely on the physical incarnation of credit and debit cards, but still use the payment rails that have been used for banking card payments for decades. Also these “pay by face” systems rely on large face databases to provide recognition, that can become a honeypot for hackers.