AI Music: Exploring Artistic Trends and Deep Learning in Music Creation

The integration of Artificial Intelligence (AI) into music creation has seen exponential growth, fueled by advancements in text-to-music generators. While ethical and copyright concerns have arisen, the fundamental question remains: How are artists leveraging AI to create music? This article delves into this question by examining a comprehensive collection of AI music artworks, analyzing them both quantitatively and qualitatively, and framing the findings within a historical and cultural context.

To understand the landscape of AI in music, a diverse range of AI music works were collected from festivals, hackathons, artist websites, and social networks. The criteria for inclusion focused on artists who self-identify as such and engage in artistic discourse, excluding casual creators. The search was guided by keywords such as 'artificial intelligence,' 'deep learning,' and 'music,' with a focus on works utilizing deep learning technologies. The data collected for each artwork included project name, artist name, year of publication, release format, and AI usage, categorized into AI-composition, co-composition, sound design, lyrics generation, and translation.

The quantitative analysis revealed several key trends in AI music. Out of 337 music artworks, the majority were singles, performances, and albums. AI is predominantly used as a co-composition and sound design tool, with significant usage in lyrics generation. The data also indicates a steady growth in AI music production since 2017, although the volume of works has stabilized in recent years. Notably, the use of AI for lyrics generation has been decreasing since 2021, primarily in singles.

The qualitative analysis explored the artistic agency and sonic characteristics of AI music. Most artists seek to retain creative control by using AI as a co-composition or sound design tool rather than relying on full AI-composition. This approach allows artists to steer and assemble models towards their creative goals. However, some artists intentionally explore the absence of artistic agency, critically embracing AI to challenge prevailing narratives. The sound of AI music varies widely, ranging from nearly indistinguishable from non-AI music to embracing the flaws and aesthetics of AI. Some artists cultivate the uncanny, blending familiar musical elements with AI artifacts to create a unique, AI-driven sound.

A significant trend is the use of generative AI as an artistic medium for releasing AI voices and generative albums. Artists like Holly Herndon and Grimes have released AI models of their voices, enabling collaborations and raising questions about digital identity and authorship. SENAIDA released Kunst Kaputt, a collaborative generative AI album where listeners can remix tracks and share profits, challenging traditional distribution models.

AI is increasingly influencing mainstream music, as seen in controversies surrounding AI-generated vocals mimicking popular artists. Despite the controversies, AI played a crucial role in the Grammy-recognized production of Now and Then by The Beatles, where AI techniques restored John Lennon's voice. AI-generated music is also attracting substantial audiences online, with projects like The Velvet Sundown and Brainrot Italianini gaining millions of streams and views.

Many AI-based music works are designed for live performances, often featuring interactive AI instruments and question-answer formats. Artists are exploring novel computer-based interfaces and the intersection of AI and acoustic music. AI DJ sets are also gaining popularity, showcasing AI's potential in live music settings.

AI music is exhibited as installations, exploring AI's potential for real-time music generation and interactive sound design. These installations often react dynamically to audience presence, integrating generative visuals and soundscapes. A notable portion of these installations are online, providing continuous livestreams of AI-generated content.

While AI offers exciting possibilities for music creation, challenges remain in addressing ethical concerns, copyright issues, and ensuring artistic agency. Future research should focus on developing more controllable AI models, improving methodologies, and creating open datasets. Further exploration of the uncanny aesthetic, multi-genre music, and novel uses of AI as an artistic medium will also shape the future of AI music.

AI is transforming the landscape of music creation, offering artists new tools and mediums for expression. From co-composition and sound design to live performances and installations, AI is pushing the boundaries of music and challenging traditional notions of authorship and creativity. As AI technology continues to evolve, its impact on music will only deepen, paving the way for new artistic trends and innovative collaborations.