No card, no phone, just you: Pay-by-face expands in Korea

A customer makes a payment using facial recognition software developed by Toss at a GS25 convenience store in Seoul, Feb. 12. Courtesy of GS Retail

Facial recognition technology, once used primarily for identity verification in Korea —such as unlocking smartphones or clearing airport immigration — is now making inroads into the payments sector, industry officials said Sunday.

Payments made solely through facial recognition, or "pay-by-face," are being adopted rapidly as financial companies compete to gain an early foothold in next-generation financial services.

Shinhan Card was the first to act, launching a pilot program in 2019 at its headquarters' cafeteria and affiliated cafes. The program was later expanded to a convenience store on the Hanyang University campus and at Homeplus locations.

While the initiative was bold, the requirement to register one's face in person at banks proved inconvenient. Public discomfort with facial recognition technology was also considerably higher at the time.

Today, Shinhan's service remains available on a limited basis, primarily at the company’s headquarters and at select convenience stores and supermarkets.

A turning point came last year when fintech giants Naver Pay and Toss entered the market, aiming to extend their dominant online market share into offline retail channels.

Their core customer base — tech-savvy young people — closely aligns with the target audience for pay-by-face. Moreover, the face registration and consent process is handled entirely within the app, providing users with a convenient onboarding experience.

In March last year, Naver Pay started operating the system on university campuses. The upcoming launch of its proprietary payment terminal in the fourth quarter is expected to accelerate adoption further.

"It’s a service built on artificial intelligence (AI) technology that delivers a level of convenience users won’t want to give up once they've experienced it, while also featuring an advanced security system with built-in fraud detection," said Lee Seung-bae, chief technology officer of Naver Financial.

A student uses Naver Pay’s facial recognition payment service at a cafeteria on Kyung Hee University's campus in Seoul, March 2024. Courtesy of Naver Pay.

The most notable player in the space is Toss, which began its service this February.

Toss addressed one of the most significant barriers to adoption: the need for merchants to install facial recognition-enabled payment terminals.

Prior to the full-scale launch, the company led the way by distributing its own terminals and offering subsidies to encourage uptake.

According to the firm, about 160,000 merchants nationwide are now using Toss terminals, which support facial payment functionality if activated. Currently, more than 20,000 stores across Seoul, including major convenience store chains and gas stations, have enabled this feature.

As the number of merchants using Toss terminals continues to grow, their adoption is expected to accelerate further.

Other firms are gearing up to introduce similar services.

The most recent example came on Wednesday, when financial authorities approved a facial recognition payment solution from Lotte Card during their regular meeting. The system will allow users to make purchases at airports using biometric data registered with the Korea Airports Corp.

However, despite the rapid growth, it remains uncertain whether the industry can scale efficiently. Because facial recognition data is highly sensitive, financial institutions cannot share it, leading to limited compatibility between different providers.

"The market is likely to evolve into a winner-takes-all scenario as merchants are unlikely to install multiple payment terminals from different providers," an industry official said.

Concerns surrounding the collection and use of facial data also remain a long-term challenge. Once compromised, facial information is difficult to restore and is particularly susceptible to forgery or tampering.

"Under current financial regulations, companies implementing such technology may be held liable for any resulting damages," the SNU AI Policy Initiative said. "This places the burden on businesses to invest heavily in developing highly accurate facial recognition models to ensure reliability and minimize risk."