SKT Ad Features Jang Wonyoung's Friend Nasu-a as a 'Virtual Human'

A.D.TV Selects Virtual Human as Main Advertising Model
Computer Graphics + AI Voice

Advertising cut for 'A. tv' featuring SKT exclusive models Jang Wonyoung and virtual human Nasua appearing as friends. Photo by SKT

[Asia Economy Reporter Oh Su-yeon] SK Telecom announced on the 26th that it has selected a ‘Virtual Human’ featuring its proprietary AI voice technology as the main advertising model for its flagship AI service ‘A. (A-dot).’

This is the first time SKT has adopted a virtual human as an advertising model in its approximately 16 years of producing over 1,300 TV commercials. It is also the first case among domestic mobile carriers. The virtual human chosen as the advertising model is named ‘Nasu-a (SUA).’ It is Korea’s first real-time interactive AI-type virtual human, created using proprietary technology by Onmind, a 3D virtual human development company.

Onmind is a company selected as the first investment target by SK Square, an investment firm spun off from SKT, in November last year. Nasu-a’s appearance in the SKT commercial marks the first fruit of the business partnership between the two companies. In the advertisement, Nasu-a appears as a friend of SKT’s exclusive model, top idol star Jang Won-young, introducing the main features of A-dot’s new service, ‘A-dot TV (A. tv).’

‘A-dot TV’ recommends channels and programs based on customers’ viewing history and preferences, allowing users to enjoy content for free without advertisements. It consists of about 10 channels including movies, dramas, and sports. It also has a feature that automatically downloads preferred content to smartphones, enabling users to watch personalized recommended content immediately in the ‘My Channel’ menu.

SKT explained, “To highlight the AI elements of A-dot TV, we judged that a setting featuring both a real human model and a virtual human simultaneously was appropriate, which led to the selection of Nasu-a.”

For the A-dot TV commercial, the technical team focused their expertise on computer graphics work to create natural acting between the real person and the virtual human. They extracted facial movement information based on the body model (real person) footage of Nasu-a, matched it to Nasu-a’s face at Onmind, and quickly synthesized the edited footage to shorten production time.

SKT applied its proprietary AI-based voice synthesis technology to realize an AI voice suitable for Nasu-a’s bright and clear image. SKT’s AI-based voice synthesis technology has previously been applied to KBS radio, NUGU, T-map celebrity voice services, and recently to the voice of the A-dot service character.

Until now, most commercials featuring virtual humans were produced with voice actor dubbing, but this commercial is the first to use AI voice created with proprietary technology for dialogue, making it the first case where everything from the face to the voice of a virtual human is created by AI and applied in advertising.

Kim Ji-hyung, head of SKT’s Integrated Marketing Strategy, said, “We have strived to highlight A-dot’s distinctive image, different from other AI services, by utilizing virtual human technology and AI voice synthesis technology,” adding, “We will continue to integrate various technologies such as AI into the company’s services and advertisements in the future.”