Musical AI: Posthumous preservation or destruction?

Celia Cheng / V Mag at UVA

The world of music is no stranger to a suddenly new album or single after the death of a beloved artist. From The New and Untouchable Death Row Records releasing 2pac’s nearly complete “The Don Killuminati” only two months after his death to Island Records dropping “Lioness: Hidden Treasures” full of unreleased Amy Winehouse songs, the posthumous release can be seen across genres. 

In the past, producers were only capable of so much when it came to fine tuning raw, unfinished tracks. However, as artificial intelligence technology continues major improvements and enters the world of arts and entertainment, artistic recreations and re-releases have grown even more accessible. Now, many outlets are debating the ethicality of using AI to “bring back the dead.

This past November, The Beatles released their first “new” song “Now and Then.” Using AI and sound technology, Paul McCartney and Ringo Starr, as well as their producers, were able to separate parts of the song and overlay new recordings. Although working with the original tape and separating the tracks was quite difficult at first, the new technology was able to “extricate John’s voice from a ropey little bit of cassette,” Paul McCartney said. 

Because the tracks were unpolished and merely bits of a potential song, a type of “audio learning” AI was used to finetune and complete what a regular soundboard could not achieve. The samples of John Lennon and George Harrison blended seamlessly with new tracks of McCartney and Starr.  

The Beatles’ new production received mixed reviews from both the media and fans. Though most listeners were excited by the return of The Beatles, conflicting opinions arose regarding the use of AI. Some listeners praised the innovation, claiming The Beatles rose to fame because of their daring use of new technologies. Others feared what this could mean for future dead artists and their music. The worry stems from the potential theft of creative jobs and extends to the risk of an artist’s wholeness.

In a blog post for Wandering Tunes, author Selin Gökova makes an important distinction between intended versus unintended posthumous releases. Gökova identifies “intended” works as those the artist had planned to release before their untimely death, while “unintended” pieces are unreleased tracks that the artist recorded but deemed unfit for release. These works aim to either pay tribute to the deceased or make money and meet demands of record labels, producers and estate holders. 

“The posthumous albums that are mostly considered to be unethical are the ones that are released with no intention other than making money out of a dead artist,” Gökova writes.

Releasing finished music may pay respect to the artist, like with Queen’s “Made in Heaven,” an album that lead singer Freddy Mercury wrote knowing he probably would never see its release. Yet some critics believe that continuously releasing music after the artist’s death is nothing more than a profit grab.The issue often arises when the artist intentionally chose not to release the music. In this case, the final product serves largely to increase the wealth of those producing. 

Some creators are going beyond the use of voice enhancement and are using artificial intelligence to mimic the vocals of other living artists. A ghostwriter used generative AI to write the song “Heart on My Sleeve” that was written and recorded with original vocals, and then overlaid with a filter to match that of The Weeknd and Drake. 

After being submitted for a Grammy, however, the song was met with a response from an official representative stating “only human creators” are eligible for consideration. The unapologetic copy and imitation of voices, dead or alive, opens up an entire debate of moral use, profit and procedure. 

Late artists like David Bowie and Michael Jackson have records upon records of unfinished or unreleased music — but the question remains whether AI development of these records is ethical or even culturally necessary. It is also important to note that posthumous production is taking on a whole new form with the involvement of AI technologies. Not only is AI being used as a crutch to assist record development, it is being used to flat out replicate voices, which has never been done before. 

If music is produced from only scraps of what could have been, who’s to say it honors their legacy instead of takes from it. Forbes has described AI’s entrance into music as a “transformative wave,” but only time can tell what the exact transformation will be.