Ushering in a new music era: How generative AI is being used for song production

The music industry has always been a hotbed of innovation, continuously stretching the limits of creativity and technology. In recent years, generative AI has emerged as a transformative force, offering the potential to revolutionise the production of music.

Earlier this year, we witnessed the launch of Suno, hailed as the ‘ChatGPT for music,’ utilising generative AI to produce human-like songs instantly from brief text cues. The platform demonstrates the remarkable fusion of AI-generated lyrics, vocals, and music, delivering a complete song experience with just a click.

As AI continues its expansion into entertainment, it's evident that its influence cannot be restrained. Generative AI possesses the power to reshape the musical landscape as we know it, highlighting the symbiotic evolution of technology and musical expression.

With AI-generated platforms successfully creating audiovisual artworks, language learning models, now music, and everything that was once solely associated with human creativity, it is worth our while to stay updated with its latest potential.

"Music is copyrighted with strict laws; however, it is very easy to steal and use it for free,” says Alex Mubert, co-founder of a pioneering AI music startup based in the UAE.

“Some AI startups exploit this vulnerability for their benefit and disregard the ethical considerations of the datasets. The same holds true for images, videos, and other creative content."

"We hope that measures will be implemented on a technical level (such as fingerprints or watermarks) to prohibit such misuse, allowing creators to earn money from AI training on their art," he says.

With a mission to revolutionise music creation, the Russian music producer and entrepreneur introduced a groundbreaking concept back in 2017: real-time generative AI music. This occurred at a time when generative AI wasn’t being utilised for music production as it is today.

“When we initially launched Mubert, it marked the inception of one of the first-ever live music generators, a pioneering AI music platform. Prior to our innovation, there were some experiments, such as those conducted by Brian Eno, but none had achieved real-time functionality, nor had any placed musicians at the core of the platform, and made it accessible across iOS, Android, and web interfaces.”

The essence of his innovation lies in its fusion of AI technology with human musicianship. Unlike previous experiments in AI-generated music, Mubert argues that his platform places human creativity at its core.

"The reality is, without humans, AI cannot exist. AI learns from datasets created by humans,” he adds. “We compensate every musician who contributes data to our platform. Currently, our dataset comprises three million ethically sourced exclusive sounds, with ongoing expansion."

"We intend to launch a royalty distribution platform to ensure musicians are compensated each time their sounds generate revenue for us," says Mubert.

Central to Mubert's vision is the belief that AI and human creativity are not mutually exclusive but complementary forces. Every sound generated by the platform is sourced from human musicians, ensuring that creativity remains at the forefront.

"When you tap 'Play’, it begins to generate music tailored exclusively for you. Each listener enjoys personalised streams. The technology behind it mirrors the creative process of musicians but is fully automated and personalised.”

“AI algorithms analyse sounds, generate new ones, arrange them, blend sounds, mix, master, and stream music. Additionally, we gather text prompts, settings, and user interactions to further personalise the experience,” he explains.

From producing royalty-free music tailored to specific use cases to personalised music experiences that adapt to individual moods and habits, AI holds immense promise for reshaping the entertainment landscape. However, with great innovation comes ethical considerations.

Mubert acknowledges the copyright implications inherent in AI-generated music and emphasises the importance of respecting artists' intellectual property rights.

"We empower our artists to select the licence for their content, allowing them to later specify the types of companies, apps, and other restrictions for their sounds' usage. This approach grants them the freedom to ethically and transparently monetise their content, ensuring they are compensated each time their music is utilised,” he says.

As AI continues to permeate the creative process, he envisions a collaborative future where human ingenuity and machine intelligence converge. "From a product standpoint, AI has long been integrated into music production and will continue to delve deeper into the creative process.”

“Artists can start creating tracks using AI and then infuse them with their unique creativity. They can also perform vocals over AI-generated beats and collaborate with renowned artists by utilising AI-driven replicas,” he adds.

Amidst the promise of AI-driven creativity, there is a growing concern that it could foster the homogenisation of creative output. Is there a risk that AI might standardise music, resulting in a landscape where all compositions begin to sound alike?

"That's accurate. The diffusion of language models often yields homogeneous outcomes. Nowadays, virtually all content on social media platforms seems to be AI-generated, which eventually makes it all look the same. This phenomenon could extend to music as well,” says Mubert.

“This is why our focus lies in exploring generative patterns, resynthesis, and the amalgamation of models to foster the emergence of new genres, song structures, and musical approaches.”

For creativity to flourish, artists also require the freedom and opportunity to experiment and push their boundaries without the pressure of always delivering immediate returns on investment (ROI). Can formulaic music generated through AI, which might promise certain ROI, hinder experimentation?

“For sure, it is a possibility and cause for concern. Everyone in the world thinks music is as cheap as $10 per month. Musicians used to be quite rich in the times of vinyl-selling copies. Now they can only earn from royalties (close to zero dollars), selling merch or live gigs,” says Mubert. “However, we stand for musicians in this battle and will do our best to protect their best interests.”

In the Middle East, Mubert sees both challenges and opportunities in implementing AI-driven solutions for entertainment. While the region's music industry is on the rise, there exists a burgeoning youth culture eager to explore new avenues of expression.

Through initiatives like preserving local music heritage and empowering local creators, Mubert aims to catalyse a new era of musical innovation in the region. "The music industry in the Middle East is still emerging, with a relatively smaller market."

"However, there is a noticeable trend among the youth to enhance the music culture, presenting ample opportunities for them to pursue careers as musicians. AI is already playing a role in this landscape, offering assistance and innovation,” says Mubert.

Slated to speak at the Integrate Middle East Summit 2024 taking place from May 21-23 taking place at the Dubai World Centre, he shares his vision for giving the local music scene an AI boost.

"My dream is to preserve local heritage by collecting instruments, melodies, rhythms, beats, and loops, and creating AI-driven public music accessible to everyone."

“This platform would serve as a futuristic hub for preserving heritage and we are constantly searching for niche styles, instruments, and rare melodies to put it all together. I look forward to discussing this further at the summit," he signs off.

somya@khaleejtimes.com

ALSO READ:

• From code to conversation: Can AI teach robots Arabic?

• 19 chapters in 19 hours: This UAE resident co-authored her memoir with ChatGPT

• Can AI make cinema a visual treat without compromising on storytelling?

• Entertainment
• Newsletters
• Dubai