Singer Randy Travis has just released his first song since suffering a stroke in 2013. Titled “Where That Came From”, the song has a twist that will not leave anyone indifferent: the vocal performance is done by an artificial intelligence software and a substitute singer.
Reminiscent of the folk and sentimental style that catapulted Travis to fame, “Where That Came From” has been produced using an AI model trained with 42 isolated vocal recordings of the singer, as reported by The Verge. Under the supervision of Travis and his producer Kyle Lehning, country singer James DuPre provided the vocals that the AI would transform into Travis’ voice.
The result is a smooth melody that captures Travis’ characteristic relaxed style, staying true to his baritone tone. Although it is not classified as one of Travis’ best songs, “Where That Came From” has received a fairly favorable reception from the audience. The song is available on YouTube, Apple Music, and Spotify platforms.
This experiment conducted by Warner Music demonstrates the potential of artificial intelligence in music creation, highlighting the ability of these tools to replicate and honor the style of iconic artists like Travis. Despite the criticism that may arise about the authenticity of AI-generated music, many people in the music industry argue that this approach strengthens creativity and innovation in the music industry.
Cris Lacy, co-president of Warner Music Nashville, has expressed her support for the use of AI to recreate Travis’s voice, calling it a significant advancement in music production, and rejecting the use of AI voice cloning websites, which use singers’ voices without the permission of record labels.
Precisely, this trend can trigger a series of legal problems, especially with the recent approval of the ELVIS Law in Tennessee (United States), which grants record labels the ability to take legal action against unauthorized use of artists’ voices.
