KT Restores Shin Hae-chul's Voice with AI "Learning Speech Patterns and Intonation"

Production of 3 Radio Broadcast Episodes

[Asia Economy Reporter Cha Min-young] KT announced on the 12th that it has restored the voice of the late singer Shin Hae-chul, known as the 'Ma-wang,' using artificial intelligence (AI) technology.

KT utilized its self-developed personalized text-to-speech technology (P-TTS) to train AI with radio broadcast data from "Shin Hae-chul's Ghost Station," which Shin Hae-chul hosted for 11 years from 2001 to 2012.

The P-TTS technology uses deep learning to synthesize a specific person's voice from a short amount of voice data.

KT explained, "Our AI goes beyond simply pronouncing sentences; it learns speech patterns and intonation to reproduce Shin Hae-chul's voice as if he were speaking naturally."

The realized voice of Shin Hae-chul can be experienced in the radio broadcast-style content titled "AI DJ, Meeting with Shin Hae-chul."

KT recreated Shin Hae-chul expressing candid opinions on popular music policies, focusing on the difficulties faced by indie bands during the COVID-19 era.

KT will provide three pieces of content through Giga Genie until October 7. Users can listen by requesting "Play Shin Hae-chul's voice" on Giga Genie. Customers who do not use Giga Genie can check the content on KT's YouTube channel, which will be released on the 17th.

Writer Bae Soon-tak, who was the main writer for "Bae Cheol-soo's Music Camp," where Shin Hae-chul also participated, and currently hosts MBC Radio's "Bae Soon-tak's B-Side," also took part in "AI DJ, Meeting with Shin Hae-chul."

This content was produced with the assistance of AI translation services.