Close
Digital Resurrection: The Ethics of Using Generative AI to Honor and/or Exploit Dead Talent

Digital Resurrection: The Ethics of Using Generative AI to Honor and/or Exploit Dead Talent

Although the music industry has been exploiting dead artists by releasing their unfinished material for quick cash since the 70s, with shoddy posthumous releases like Crash Landing by Jimi Hendrix and An American Prayer by Jim Morrison, the ethicality surrounding record labels taking advantage of dead artists has become more relevant than ever as of late. While posthumous projects like Bad Vibes Forever from the late XXXTENTACION and The Party Never Ends from the late Juice WRLD have previously exemplified the lengths to which record labels will go to make money off of dead talent, lately, these practices have been sidelined by a whole new threat to the legacy of artists: Artificial Intelligence. On the surface, using AI to create a song that mimics an actual musician’s style, and voice seems like a ridiculous proposition with very little appeal. A song generated by an AI model intended to copy Taylor Swift, for example, is unlikely to captivate even her most hardcore fans, because as convincing as the AI model may be, it just isn’t Taylor Swift. For that reason, and many more, songs created by human artists with dedicated fanbases continue to dominate the charts. However, when it comes to dead artists, who obviously can’t put out music themselves anymore, and have run out of demos and outtakes for their labels to repackage, there seems to be at least a little support for the idea of resurrecting these musicians using new technology.

In 2021, a project called “The Lost Tapes of the 27 Club” was kicked off by an organization called Over The Bridge. This project aimed to honor members of The 27 Club—an informal collective of celebrities who died at the exact age of 27—and convey how tragic and untimely their losses were by releasing music inspired by their work. Dozens of existing songs from musicians like Kurt Cobain and Amy Winehouse were fed to an AI program, which used that data to create brand new songs in their respective styles. Nowadays, anyone can ask an AI program to create a piece of music inspired by any given artist, and they’ll have a brand-new song within seconds. However, in 2021, over a year before ChatGPT was made available to the public, this was exciting technology. As such, the positive press and numerous awards the project received were hardly surprising. However, even at the time, many people were skeptical of the ethics of this new technology, with many YouTube comments on these videos pointing out that the computer simply couldn’t replicate the soul of Amy Winehouse’s lyrics and singing, or the specificity of Jimi Hendrix’s legendary guitar shredding.

Controversy surrounding the use of AI to resurrect late musicians popped up again a couple years ago in the wake of the ongoing feud between rappers Drake and Kendrick Lamar. In April of 2024, Drake posted a diss track titled “Taylor Made Freestyle” to his social media platforms, which featured some bars from Drake himself, as well as verses generated by AI models based off Snoop Dogg and 2Pac. The idea of having some of Kendrick’s greatest inspirations, dead and alive, belittle him on this song might’ve seemed brilliant to Drake, but at this point in time, people were already aware of how convincingly AI could replicate real human voices, and how frequently people employed them to do so for both comedic and practical purposes. For that reason, many of the responses to this move were focused on how crumby and cringy it was, rather than how clever and sinister it was. It’s worth noting that Drake dropped this track shortly after his first diss against Kendrick, “Push Ups,” and before Kendrick put out his initial responses to the beef in the form of “Euphoria” and “6:16 in LA.” Had Drake used the AI voices of Snoop Dogg and 2Pac on a song like “Family Matters,” which dropped right after Kendrick’s initial responses, it’s likely that he would face even more backlash, as using generative AI like that, on a song that’s supposed to make a statement, can sour the message.

Overall, the music industry has largely been shielded from the threat of generative AI, at least as it relates to exploiting dead musicians by having a computer generate brand new songs based off existing material. Although “The Lost Tapes of the 27 Club” might’ve enamored some fans of artists like Kurt Cobain and Amy Winehouse at the time, people have largely grown uninterested in the premise it introduced, potentially because the field is so saturated that it takes away the novelty of the idea, and also because it just feels wrong. As such, no AI generated songs attributed to dead musicians have risen to the top of the charts. When it comes to Hollywood, on the other hand, generative AI plays a very different role and has very different implications.

In 2025, actor Val Kilmer, best known for his role as Tom Kazansky in the Top Gun films, passed away after a long battle with throat cancer that made it so he could rarely come to set on a movie he had been working on, As Deep as the Grave. After he passed, with his role in the movie not being completely fulfilled, the director saw it fit to combine archival footage with the output of generative AI programs to create an AI version of Kilmer to finish the job. It’s hard to tell how well this decision will play out practically, since As Deep as the Grave doesn’t release until later this year, but reception to the move has been incredibly mixed, with some people defending it, seeing as Kilmer’s daughter and family approved of the decision, with others criticizing it, because it’s disrespectful and disingenuous to portray Kilmer in this light.

Years before generative AI took off, becoming one of the largest controversies and talking points of the 2020s, James Wan used CGI to wrap up scenes featuring Paul Walker in Furious 7, after he died in a tragic car accident before filming was completed. There was much less backlash to this decision compared to the controversy surrounding Kilmer, which does beg the question: Are people only mad at AI because it’s a new and confusing technology? Will all this discourse simmer down within a matter of years? Personally, I don’t think so. Over the past few years, there has been some overblown backlash to various forms and uses of generative AI. In 2023, The Beatles released their final song with the help of generative AI, which was implemented very sparingly, only to isolate John Lennon’s vocals from a shoddy recording. Still “Now and Then” faced a great deal of backlash for utilizing AI in any shape or form. In instances like these, I can understand why some people might be fed up with the way generative AI is labeled as a malevolent force without any nuisance. However, I’m largely opposed to the concept of dead talent being digitally resurrected through the use of AI, because although late artists and movie stars have been exploited by executives since long before the 2020s, the inability for a computer to organically process emotion and convey nuisance makes projects like “The Lost Tapes of the 27 Club” and As Deep as the Grave seem even more disrespectful to their legacies.

Leave a Reply

Your email address will not be published. Required fields are marked *

Close