More than 60 years after their debut single, the Beatles have released a new recording, “Now and Then,” advertised as the last ever Beatles song. With classic Beatles symmetry, their first release, “Love Me Do,” serves as the B side for this last song. Not as classic, however, was their use of newly created artificial intelligence (AI) to create the track.
In the late 1970s, John Lennon wrote and performed a demo of “Now and Then” on his cassette recorder, which was given to the surviving Beatles members approximately two decades later by his widow, Yoko Ono. While working on “The Beatles Anthology” retrospective project, the group attempted to use the vocals from the demo but encountered audio issues with the recording. The cassette tape Lennon recorded made for a messy demo: It was scratchy with a persistent electric buzz, the TV could be heard in the background and Lennon’s voice was on the same track as his piano — with one often drowning out the other. There was little they could do with the technology of the time, so the band abandoned the song.
Using technology made possible due to recent advances made by film director Peter Jackson and his team — developed while creating the documentary series “The Beatles: Get Back” — Paul McCartney and Ringo Starr were able to isolate Lennon’s vocal track from the rest of the original recording. The machine-assisted learning (MAL) technology developed by Jackson’s team could distinguish between different instruments and voices. The MAL machine “allows us to take any soundtrack and split all the different components into separate tracks,” Jackson said in a new short film about the making of “Now and Then.” The technology separated Lennon’s vocals from the piano parts on the “Now and Then” demo, and “there it was, John’s voice, crystal clear,” McCartney said in the film.
Controversy in the Music Industry
In the Beatles’ case, because Ono and the remaining members of the band participated in the project and all relevant parties consented, the band’s use of AI may well be the least controversial issue in the music industry. However, putting aside the debate over the quality of the song or whether it is a “real” Beatles release, the technology used to create the track has resulted in controversy. In a recent appearance on “The Tonight Show,” singer-songwriter Sheryl Crow argued that a machine creates AI music and, therefore, that music is without soul.
AI is everywhere. It has long been behind the scenes functioning in largely uncontroversial ways. AI processes photos on smartphones and provides wording prompts when texting — and it is an emerging tool for making music. Unlike other useful applications such as auto-tune, AI can replace writers, artists and musicians. Will it? Unchecked, the answer is undoubtedly yes; it already is.
Case in point: During that same “The Tonight Show” appearance, Crow told host Jimmy Fallon that she had met a young songwriter who had produced demos that the songwriter intended to pitch to established singers. But there was a problem: The upstart composer needed a male singer to perform on one of the demos. Rather than hire a singer, Crow said, the composer paid $5 to have an AI application reproduce the sound of singer-songwriter John Mayer singing over her demo. It is fair to assume that Mayer had no say in this creation and did not receive any compensation. While not a wide release, it still seems a highly problematic use of a talent’s protectible rights.
In an even more troubling example, a song created by a user through the AI music composition platform Amper Music, “Heart on My Sleeve,” featured AI versions of musical artists Drake and The Weeknd. The track, which went viral, was uploaded to well-known streaming services but was quickly removed following copyright infringement claims by Universal Music Group. The song was also submitted for consideration for a Grammy Award but held ineligible for various reasons, including that it was not generally commercially available.
SAG-AFTRA Tentative Deal Includes AI Provisions
These concerns are not unique to the music industry. The SAG-AFTRA union, which represents tens of thousands of actors, was, until recently, on strike for about four months. One of the main issues was the use of AI by producers to create digital replicas of talent without informed consent and fair compensation. For example, the likenesses of both actor Tom Hanks and TV personality Gayle King were used in advertisements that were not authorized by either.
The final details of the tentative agreement have not yet been released. However, the Alliance of Motion Picture and Television Producers — the group that represented studios, streaming services and production companies in the negotiations — said in a statement that the Nov. 8 deal “represents a new paradigm” that “gives SAG-AFTRA the biggest contract-on-contract gains in the history of the union, including the largest increase in minimum wages in the last 40 years; a brand new residual for streaming programs; extensive consent and compensation protections in the use of artificial intelligence; and sizable contract increases on items across the board.”
SAG-AFTRA was concerned that studios could use AI to reanimate actors who had passed away or to create a digital Frankenstein out of actors’ body parts. In the negotiations, the union secured a requirement that if a Frankenstein actor contains recognizable features of real-life actors, studios must get permission from those actors, who also must be paid for the performance. “If you’re using Brad Pitt’s smile and Jennifer Aniston’s eyes, both would have a right of consent,” said Duncan Crabtree-Ireland, the union’s chief negotiator, in an interview with Variety.
Executive Order Aims to Balance AI Benefits, Risks
These issues, as well as security and other concerns, resulted in President Biden issuing a first-of-its-kind executive order on Oct. 30, which seeks to balance the benefits of AI with its inherent risks. The executive order aims to establish new standards for AI safety and security, protect Americans’ privacy, advance equity and civil rights, stand up for consumers and workers, promote innovation and competition, and advance American leadership worldwide. The order also establishes standards and practices for detecting AI-generated content and authenticating the creators of content.
This latest development is a start, but Congress will have to act swiftly and decisively to control the use of AI or, at a minimum, to ensure the consuming public is informed when AI is used in creative endeavors. Historically, the law is years behind new technology. In this instance, it may be too long to wait.
Meet the Author
Randy M. Friedberg
Randy Friedberg is a member of the firm’s business department. A seasoned attorney, Randy has advised clients and businesses of all sizes for over 30 years on trademark and copyright law, unfair competition, trade secrets, advertising, internet and cyber issues, rights of privacy and publicity, entertainment law, and general corporate and litigation matters. With a particular focus on entertainment and media companies, including advertising and marketing agencies, talent, artists and art galleries, Randy often provides counsel on protection, licensing and enforcement of intellectual property rights – routinely finding creative solutions to highly technical and complex issues. He has represented a broad range of industries, including consumer goods and services, computer software and hardware, pharmaceutical and biomedical products, non-profits, energy and financial services.
212.404.0634