As generative AI sends shockwaves throughout the music industry, BBDO New York interactive producer Marcelo Hanta-Davis reflects on some lessons learned from a recent project.
Last month, a long roster of musicians including Billie Eilish, Stevie Wonder and R.E.M. signed an open letter calling for an end to the unethical use of AI within the music industry. The letter argued that the unchecked use of AI – using an artists’ songs without consent to train a model, for example – can damage an artist’s image and livelihood.
As an interactive producer at BBDO with experience working on AI music industry projects, I consider this a watershed moment and a much-needed call to action. This letter marks the first-time artists of this caliber have come together to call out the predatory use of AI within their industry. But since musicians are not unionized, it’s harder for them to collectively push for change. In my view, we need legislation on a national scale in the US to protect artists from being adversely affected by AI.
The adverse effects that AI impersonations can have on music have become increasingly apparent. On April 13th, an early demo version of Drake’s diss track, “Push Ups,” aimed at Kendrick Lamar, was leaked to the public. Fans quickly speculated that the demo was an AI-generated song, a claim that was later disproved when Drake released a final version. On April 15th, Kendrick Lamar seemed to drop his own diss track, but this was later revealed to have been created by AI. The confusing back-and-forth led fans to take to social media and ask for an end to AI impersonations without the artist’s consent. Clearly, unrestrained AI in music is bad for everyone, whether you are an artist, fan or part of the music rights business.
Drake’s latest use of AI to rap in the voice of Tupac and Snoop Dogg highlights an even more significant moment for AI in Hip-Hop, marking the first time that a prominent artist has used AI to recreate the voice of real people in a song. Now, Tupac’s estate is suing Drake for the usage. Is it OK to do AI impersonations in music when they are comedic? Would that count as a parody and fair use in this situation? As questions like these continue to arise, it’s becoming more apparent that clear guidelines are necessary.
We should not rely only on legislation. Marketers need to set a standard for ethically using AI in music. At BBDO this past year, we saw the possibilities of what can happen when artists, brands and tech companies work together to embrace AI responsibly. Over the past decade, we’ve worked with Bacardi on a program called Music Liberates Music (MLM) that empowers up-and-coming artists with new tech. Through our time working on this project, we noticed that these artists often don’t have access to high-quality production. That led us to an idea: What if AI could help to provide a solution?
We convinced Grammy-winning producer Boi-1da to train an AI model on his beats as part of our most recent iteration of MLM. In collaboration with ethical AI companies Tuney and Audioshake, we created a tool which allows artists to upgrade their demos to a Boi-1da-style song with the help of AI. Our interface would analyze the artist’s demo, insert Boi-1da instrumentals, and output a new song – all while keeping the artists’ vocals intact. All the AI training was done tracks that were provided consensually.
We had three main takeaways from this project. First, the ethical use of AI-generated music is scalable. Second, we’re in the Wild West for this technology; the people making deals now will set a precedent for the future. And finally, making AI-generated music can be fun. When artists heard their vocals over new instrumentals, it enhanced the creative process and got them excited about the possibilities.
As generative AI has grown in popularity, many music industry leaders have focused either on leveraging the technology to quickly generate ideas or produce content more affordably. But if we zoom out, its real potential lies in speeding up the process from ideation to execution. Many people can come up with great ideas, but executing these thoughts is often the hardest part. There are limitless opportunities to turn the ideas in our brain into something real. We just need to make sure we do so ethically, consensually and transparently.
For more on the latest happenings in AI, web3 and other cutting-edge technologies, sign up for The Emerging Tech Briefing newsletter.
