Virtual Human 'Rozy' Appears on Terrestrial Radio... Naver Utilizes AI Voice Synthesis Technology

[Asia Economy Reporter Kang Nahum] Naver has created the voice of the virtual human model 'Rozy'.

On the 6th, Naver announced that it partnered with Sidus Studio X to create Rozy's artificial intelligence (AI) voice using Clova AI voice synthesis technology.

Rozy's AI voice is scheduled to be unveiled on SBS PowerFM's 'Escape at Two Cultwo Show' airing on the 8th. Rozy will appear as a special guest on the show, discussing her activities as a virtual influencer with DJ Kim Taekyoon and personally reading listener stories. This marks the first time a virtual human has appeared on a radio broadcast.

Rozy's AI voice was produced using NES (Natural End-to-end Speech Synthesis) technology, developed in-house by Naver Clova. The distinguishing feature of NES is that it can create a natural, human-like voice with only about 40 minutes of short recordings, without limitations on the field of use.

With Rozy now able to freely communicate vocally through AI voice without separate recordings, it is expected that her activities, which have so far focused on SNS and advertising, will expand significantly. Beyond being a radio DJ, Rozy is anticipated to be active in various fields requiring voice, such as docent, MC, and live commerce host.

Kim Jaemin, Head of Naver Clova Voice, said, "It is meaningful to produce the voice of Rozy, Korea's first virtual influencer, using Clova's AI technology," adding, "Naver will continue active collaboration with Rozy and keep expanding the scope of Clova AI Voice applications."