How the Music Business Used Hip-Hop to Prepare for its AI Future
The 50th anniversary of hip-hop reveals unlikely parallels between the rise of rap music and AI
This month the world is celebrating the 50th anniversary of the birth of hip-hop, the cultural phenomenon that comprises rapping, DJing, graffiti, and break dancing (which is set to make its debut at the Olympics in Paris in 2024). The occasion is a special one for me because I dedicated the start of my career to helping mainstream the culture, primarily through my work marketing artists, producing music, and most notably writing for and editing the pioneering hip-hop magazine The Source many years ago.
All four elements of hip-hop continue to thrive today, but the one most are familiar with is rap music, which now dominates the overall streaming and sales of music, according to data compiled by entertainment analytics firm Luminate.
The timing of the anniversary is tied to a party in the Bronx on Aug. 11, 1973, where DJ Kool Herc operated the turntables. But if you talk to many old school hip-hop fans who weren’t privy to the very earliest iterations of the music, they’ll likely tell you the first rap song they heard was “Rapper’s Delight,” by the Sugar Hill Gang, released in 1979. The song used a break from the R&B group Chic from their song “Good Times.”
Although The Sugarhill Gang didn’t use electronic samplers, which came into the picture years later, the popularity of “Rapper’s Delight” further fueled the growing practice of using a snippet (or breakbeat) from a vinyl record from a popular song to serve as a kind of collage piece for a new rap song.
Later, companies including Roland, Akai, and E-mu released electronic samplers that shaped the sound of ‘80s rap music produced by Larry Smith, Arthur Baker, Rick Rubin, Kurtis Mantronik, Marley Marl, and many others. By the late ’80s, as rap music found its initial footing as a viable commercial genre, sampling was finally appearing on the radar of some of the most established artists being sampled. One of the most famous early cases involved Gilbert O’Sullivan suing rapper Biz Markie due to the use of a sample from O’Sullivan’s 1972 hit song “Alone Again (Naturally)” on Biz Markie’s “Alone Again” rap song released in 1991.
The case was settled the following year, but it was just one in a mountain of lawsuits related to sampling in rap music by the musicians providing the source material for the samples. In short order, sample clearance became just as important as any other component of the production process in rap music. It turned out that recontextualizing the work of others and creating something new as a result had legal consequences that could be tracked by simply listening to the rap song using the sample.
Dawn of the Entertainment Industry’s Digital Disruption
Adding to the complications faced by both record labels and recording artists was the rise of the MP3 audio format, which allowed audio files to more easily be stored on personal computers and shared over the Internet. So while the music industry was still grappling with the rise of rap music and its legally complicated sampling foundations, a growing number of computer users began sharing music files illegally via the Internet, presenting a major threat to companies accustomed to controlling every aspect of the music purchase process.
In 1998, I attended the press conference in Manhattan organized by the world’s biggest music labels to address the sudden rise of digital music piracy. Appearing at the event were the CEOs of Sony, Universal Music, EMI, BMG, Warner Music Group, and the Recording Industry Association of America (RIAA). Collectively, the labels announced the Secure Digital Music Initiative (SDMI), which was an attempt to introduce an alternative digital format that would curtail the use of illegal MP3 files.
The effort was supported by America Online (one of the leading Internet service providers at the time), Microsoft, AT&T, IBM, and Toshiba. However, the SDMI plan never really took hold, and the following year the launch of Napster, a way station for illegal music MP3 files, was launched. Major labels were in a frenzy over the loss of control of their intellectual property, while some artists were more forward-thinking.
In 2000, I tried to do my part to smooth the way into the digital music era by launching FreeListen.com, which was featured in The New York Times as a site designed to offer artists a way to harness MP3 technology as a marketing tool. Some of the artists I convinced to participate in offering digital files on the site included The Smashing Pumpkins, The Beastie Boys, Erykah Badu, and Digital Underground, among others.
Nevertheless, the major record labels were on the warpath against MP3 technology and the growing numbers of users illegally sharing the files. One central turning point was the RIAA’s inability to legally stop Diamond Multimedia from selling its Rio portable MP3 player, which paved the way for Apple’s iPod, which arrived in 2001.
Illegal music and movie file sites still exist today, but eventually, the music industry embraced digital files. In 2001, digital music stores PressPlay and Rhapsody offered subscription-based services to serve those not engaging in illegal MP3 file sharing. Still, it was Apple’s ability to present an even easier path to legal music file purchases through the iTunes Music Store, launched in 2003, that really helped change the music industry’s attitude toward digital music and laid the foundation for how it handled streaming year later.
So what does this all have to do with artificial intelligence (AI)? If you’ve made it this far, you may have already begun to connect the dots.
As Usual, Rap Moved First to Embrace the Latest New Technology
"The genie is out of the bottle…when the industry doesn't come together and create a solution, things happen anyway,” Rob Glaser, the founder of pioneering digital audio startup Real Networks, said back when the music industry made its first real attempt to grapple with digital music in 1998. As calls for a pause in AI development arise, we’re hearing the same kind of “genie out of the bottle” language that shadowed the digital disruption of the music business (and later, Hollywood).
The storm brewing around AI, particularly as it relates to creative content, is a mix of the challenges presented in the early days of music sampling, that is, sourcing past works to create new works, and the struggle to somehow mark legal versus illegal shared digital files, or, in the case of AI, determine what is and isn’t Generative AI-produced work.
In April, an anonymous AI experimenter named Ghostwriter977 uploaded a song titled “Heart on my Sleeve” which featured convincing AI-generated vocals imitating Drake and The Weeknd singing on the track. The song, which was uploaded to YouTube, Spotify, and TikTok, was quickly removed after complaints from Universal Music Group (UMG), the parent company that distributes music produced by the real artists mimicked on the song.
“Sophisticated generative AI that's enabled by large language models, which trains on our intellectual property, violates copyright law in several ways. Companies have to obtain permission and execute a license to use copyrighted content for AI training or other purposes, and we're committed to maintaining these legal principles,” said UMG’s chief digital officer, Michael Nash, during the company’s quarterly earnings conference call on April 26.
In the past, with music samples, identifying an infringing use was straightforward. But how can an entertainment company track and reliably identify near-perfect impersonations?
“Depending on the instance, we may also employ name and likeness, voice impersonation, right of publicity protections as well,” said Nash. “And specifically, soundalikes which serve to confuse the public as to the source or origin, or which constitute a commercial appropriation of likeness in the form of a distinctive voice, those are all clearly illegal.”
In response to the Ghostwriter977 incident, UMG recently announced an initiative with YouTube meant to create a framework to address such instances.
“Our challenge and opportunity as an industry is to establish effective tools, incentives and rewards—as well as rules of the road—that enable us to limit AI’s potential downside while promoting its promising upside. If we strike the right balance, I believe AI will amplify human imagination and enrich musical creativity in extraordinary new ways,” said UMG CEO Sir Lucian Grainge on YouTube’s website.
“We are working hand-in-hand with YouTube to launch their Music AI Incubator, which is first bringing together a working group of leading UMG artists, songwriters, and producers in multiple genres. This group will explore, experiment and offer feedback on the AI-related musical tools and products they are researching.”
Will the two entertainment giants come up with an industry-leading solution? Only time will tell.
AI is Forcing the Entertainment Industry to Redefine How it Operates, Again
At least one court decision rendered this week in Washington, D.C., provided some clarity as to where UMG and YouTube’s efforts might one day lead. The court ruled that works generated solely by AI cannot be copyrighted. “This case [Thaler v. Perlmutter] presents only the question of whether a work generated autonomously by a computer system is eligible for copyright,” wrote Judge Beryl A. Howell. “In the absence of any human involvement in the creation of the work, the clear and straightforward answer is the one given by the [U.S. Copyright Office]: No.”
The indication here is that as long as a human is involved in creating a work, perhaps by also using AI tools, such work may be protectable under U.S. copyright.
“[For example] in [the case of] Sarony, the Court’s recognition of the copyrightability of a photograph rested on the fact that the human creator, not the camera, conceived of and designed the image and then used the camera to capture the image,” wrote Judge Howell.
“We are approaching new frontiers in copyright as artists put AI in their toolbox to be used in the generation of new visual and other artistic works. The increased attenuation of human creativity from the actual generation of the final work will prompt challenging questions regarding how much human input is necessary to qualify the user of an AI system as an ‘author’ of a generated work…”
This sounds like a promising perspective for the future of using AI tools that can be protected by copyright and allow the creator to profit from the work.
But in the meantime, there’s also the issue of content provenance. As Generative AI tools for producing music, video, and text improve, there will be instances in which it will be necessary to know whether or not a piece of content has been AI-generated. Such distinctions could be crucial to labeling a piece of art or text as partially human-created or solely AI-created, as indicated in Judge Howell’s case.
OpenAI, the maker of ChatGPT and DALL-E 2, has stated on its website that it cannot currently reliably differentiate between human and AI-generated text.
Similarly, in a University of Maryland Department of Computer Science research paper (PDF) published in June, a team was unable to reliably detect AI-generated text. “Current detectors of AI aren’t reliable in practical scenarios,” said Soheil Feizi, an assistant professor of computer science at the university who contributed to the research paper.
The more promising news is that a large group of companies are working on addressing every one of the aforementioned issues.
In the mid to late ‘90s, it seemed like the music and film industry might be completely destroyed by the two-headed monster of sampling and Internet file sharing. But the reality was that the entertainment industry, and the various technology and distribution businesses supporting it, needed time and experience to adjust to a completely new reality in order to devise technical and legal solutions to what seemed like a crisis moment in entertainment.
My experience with the disruption brought about by early hip-hop music, and later Internet-powered innovation, leads me to suspect that we are in familiar territory with AI. Generative AI is a tool that is already giving birth to new forms of expression that were inconceivable just a couple of years ago. When innovation arrives more rapidly than expected, it can be terrifying. But the more we familiarize ourselves with the new, usually, the less intimidating and unwieldy it becomes.
Hip-hop was a tectonic, culture-shifting innovation that was at first feared, ridiculed, and even shunned by most in mainstream culture for years before its widespread adoption. On this 50th anniversary of hip-hop, I would encourage us all to gaze upon the new landscape of AI with the same spirit of experimentation, dedication to hard questions, and search for brilliant answers that, as with hip-hop, forced people from different backgrounds across the world to look at the world with new eyes. This is the beginning of something new, but AI will mean nothing without the deeply involved hand of humanity bending the technology to our will to create things we still have yet to imagine.