AI in Music and Creative Arts: Suno, Udio

The Machines Can Create Now

In January 2024, Suno released a text-to-music model that could generate a complete, radio-ready song — lyrics, melody, instrumentation, vocals — in under 30 seconds from a simple text prompt like “upbeat Afrobeat song about summer in Lagos.” Suno v4 followed in November 2024, delivering studio-quality 44.1 kHz audio with balanced mixing and enhanced instrument layering. By May 2025, Suno v4.5 extended output to eight minutes per track and supported over 1,200 musical genres. Then came v5 and Suno Studio — a generative audio workstation blending multi-track editing with AI stem generation — alongside a $250 million Series C that valued the company at $2.45 billion on $200 million in annual revenue. Meanwhile, competitor Udio had reached comparable quality levels. A 2025 study by Deezer and Ipsos, surveying 9,000 people across eight countries, found that 97% of respondents could not identify whether tracks they heard were fully AI-generated or human-made. Researchers at NYU’s Music and Audio Research Laboratory reached a similar conclusion: when song pairs are presented randomly, listeners perform no better than chance at distinguishing AI from human music.

The same trajectory has played out in visual art. Midjourney v7, which became the default model in June 2025, produces photorealistic images with video generation capabilities. An AI-generated image by Boris Eldagsen won the Creative category of the Sony World Photography Awards in 2023 — though Eldagsen refused the award, revealing the image was AI-generated to spark debate about whether competitions were prepared for synthetic entries. DALL-E 3 and Stable Diffusion generate commercial-quality illustrations, product mockups, and concept art. Adobe’s Firefly, trained on licensed Adobe Stock content and public domain works, has been integrated directly into Photoshop and Illustrator, making AI generation a standard design tool. Runway’s Gen-4, released in 2025, produces short video clips with cinematic quality and consistent character identity across scenes.

The creative industries — music ($29.6 billion in recorded revenue in 2024, according to the IFPI, plus $35 billion or more in live performance), visual arts, film, advertising, gaming, publishing — collectively represent approximately $2.25 trillion in global economic activity. These industries employ millions of creators, from superstars to freelancers. What happens to this ecosystem when the marginal cost of producing competent creative work approaches zero?

The Economic Disruption Is Already Here

The impact on creative freelancers has been immediate and measurable. A survey by the Association of Illustrators (AOI) of nearly 7,000 illustrators found that one in three had already lost work to AI, costing them approximately $12,500 on average in wages. Research published in the Journal of Economic Behavior & Organization found that freelancers in occupations more exposed to generative AI experienced a 2% decline in the number of contracts and a 5% drop in earnings following new AI releases, with demand for substitutable skill clusters falling by up to 50% in short-term roles. Stock photography has been hit hard: Shutterstock contributors report earnings declines of 20 to 40 percent from pre-AI peaks, as the platform’s cost of revenue dropped from 42% to 39% of total revenue — meaning the company is paying contributors less even as AI-generated images flood the marketplace. Voice actors face similar disruption: ElevenLabs can clone a voice from a short audio sample of roughly 10 to 30 seconds and produce unlimited narration at a fraction of union rates.

In music, the disruption reached a symbolic milestone in November 2025 when Breaking Rust’s “Walk My Walk” — a country track created entirely with Suno — topped Billboard’s Country Digital Songs chart after accumulating over three million Spotify streams in under a month. At least six AI-generated acts have now charted on Billboard. Meanwhile, Deezer revealed that 50,000 fully AI-generated tracks are being uploaded daily to the platform, representing 34% of all daily submissions — a 400% increase from the start of 2025. Deezer became the first streaming platform to explicitly label AI-generated music and remove it from algorithmic recommendations.

Session musicians, jingle composers, production music libraries, and background score creators are seeing AI alternatives encroach on their work. YouTube launched an AI music generation feature within its Creator Music marketplace in 2025, with a “Music Assistant” tab allowing creators to generate custom instrumentals from text prompts. Google entered the space with Lyria 3 through Gemini and YouTube’s Dream Track for Shorts. A video creator who previously paid $200 for a licensed background track can now generate one for free. Epidemic Sound, which provides production music for content creators, took a different approach: rather than introducing purely AI-generated tracks, it launched “Adapt” and “Studio” — AI tools that modify and recombine its existing human-composed catalog, ensuring that original artists continue to receive compensation.

The gaming industry has been an early adopter. Indie developers use AI-generated art, music, and dialogue to produce games that would previously have required teams of artists and composers. AAA studios are more cautious — both because of quality standards and labor concerns — but internal tools for concept art generation and asset creation are widespread. EA partnered with Stability AI to co-develop generative AI tools for game development. Square Enix announced plans to use generative AI to automate 70% of its QA and debugging tasks by 2027. Ubisoft has similarly acknowledged deploying generative AI across its production pipeline.

The Legal Battleground

The legal response was swift and aggressive — and then took a dramatic turn toward settlement. In June 2024, the RIAA filed copyright infringement lawsuits against both Suno and Udio, alleging that the companies trained their models on copyrighted music without authorization. The suits sought damages of up to $150,000 per work infringed. Suno acknowledged in court filings that it trained on copyrighted material but argued this constituted fair use — the same defense used by Google Books and other transformative technology projects.

By late 2025, the landscape shifted from courtrooms to boardrooms. Universal Music Group settled with Udio in October 2025 and agreed a licensing partnership. Warner Music Group followed in November, settling its lawsuit against Suno in a deal hailed as a “landmark” for AI music. The terms differ sharply between the two companies. Udio must pivot from open-ended music generation to a “walled garden” fan engagement platform where users can remix and interact with licensed music, but nothing generated can leave the platform. Suno’s deal is less restrictive: users keep the ability to generate and download music, but Suno must train future models exclusively on licensed material, and users must pay to download tracks. Both companies have committed to retiring their current models — trained on unlicensed catalogs — and launching new, licensed-only models. Sony Music has not settled with either company, and UMG’s suit against Suno remains ongoing.

Visual artists have pursued parallel litigation with mixed results. Getty Images sued Stability AI in both the US and UK for training Stable Diffusion on Getty’s licensed images. In November 2025, the UK High Court handed down the first major ruling on AI training and copyright: Getty largely lost its copyright infringement claims, with the court rejecting the argument that scraping images for AI training constituted secondary infringement. However, Getty partially won on trademark grounds — the court found that Stable Diffusion’s reproduction of Getty watermarks in AI-generated outputs infringed Getty’s trademarks. A class-action lawsuit filed by illustrators Sarah Andersen, Kelly McKernan, and Karla Ortiz against Stability AI, Midjourney, and DeviantArt proceeded past initial motions to dismiss in August 2024, with U.S. District Judge William Orrick in the Northern District of California allowing copyright infringement claims to move to discovery. The New York Times’ lawsuit against OpenAI and Microsoft — while focused on text — advanced in April 2025 when the court allowed key copyright claims to proceed, denying OpenAI’s motions to dismiss.

The fundamental legal question remains unresolved: does training an AI model on copyrighted works constitute infringement, or is it transformative fair use? The answer is fracturing along jurisdictional lines. The UK High Court’s ruling in Getty v. Stability AI leaned toward permissiveness on training. The EU’s AI Act provides limited exceptions for text and data mining but requires opt-out mechanisms for rights holders. Japan’s copyright law has been interpreted as broadly permissive of AI training. The US, where the most consequential cases remain in litigation, has not yet produced a definitive appellate ruling. Meanwhile, the music industry’s pivot from litigation to licensing suggests that pragmatic deal-making may outpace judicial resolution.

The Philosophical Question: What Is Creativity?

Beyond economics and law, AI-generated creative work raises questions that are genuinely philosophical. Is a Suno-generated song “creative”? It was not conceived by a mind grappling with emotion, experience, or artistic intent. It is a statistical interpolation of patterns extracted from millions of human-created songs. Yet the output — melody, harmony, rhythm, lyrical content — is indistinguishable from human work to most listeners. Does the process matter if the product is equivalent?

Artists and philosophers are divided. The “tool” camp argues that AI is simply a new creative instrument, like the synthesizer or the camera — technologies that were also initially accused of threatening “real” art. David Hockney’s iPad paintings are no less art because they were made digitally. A filmmaker using Runway to generate a scene is exercising creative judgment in prompting, selecting, and editing. Creativity, in this view, lies in the intention and curation, not the manual execution. Recording Academy CEO Harvey Mason Jr. has noted that “every” songwriter and producer he knows has used Suno, often to break through creative blocks.

The “displacement” camp counters that creative tools historically augmented human capability — a camera still requires a photographer’s eye, a synthesizer still requires a musician’s composition. AI generation is qualitatively different: it can produce complete works without meaningful human creative input. When a marketing manager types “happy corporate background music, 2 minutes” into Suno and uses the first result, there is no creative act occurring. The musician whose livelihood depended on composing that piece has been eliminated, not augmented. The Deezer data underscores the scale: 50,000 AI-generated tracks per day flooding a single platform, the vast majority indistinguishable from human work.

What is increasingly clear is that the market will not wait for philosophical resolution. Businesses will use AI-generated creative content wherever it is cheaper and good enough. The question is not whether AI will displace creative work — it already is — but how society will structure the economic transition, protect displaced creators, and define the boundaries between legitimate AI use and exploitation of creative labor. The music industry’s shift from suing AI companies to licensing deals with them suggests one possible model: coexistence through compensation, rather than prohibition. Whether that model delivers meaningful value to working artists — or primarily enriches the platforms and labels negotiating the deals — remains the open question of 2026.

Frequently Asked Questions

Sources & Further Reading

• IFPI Global Music Report 2025: Global Recorded Music Revenues Grew 4.8% in 2024 — IFPI
• Record Companies Bring Landmark Cases Against Suno and Udio — RIAA
• Warner Music Settles Legal War With Suno In Landmark AI Partnership — The Hollywood Reporter
• What Do the Suno and Udio Licensing Deals Mean for the Future of AI Music? — Billboard
• Suno Raises $250M Series C at $2.45B Valuation on $200M Revenue — TechCrunch
• 50,000 AI Tracks Flood Deezer Daily; 97% of Listeners Can’t Tell the Difference — Music Business Worldwide
• Getty Images v. Stability AI: UK High Court Ruling — Mayer Brown
• Andersen v. Stability AI Copyright Case Takeaways — Copyright Alliance
• NYT Copyright Lawsuit Against OpenAI, Microsoft Can Advance — Axios
• AI-Generated Country Track Tops Billboard Chart — NME
• YouTube Rolls Out Free AI Music-Making Tool for Creators — TechCrunch
• Artists Are Losing Work, Wages, and Hope as Bosses Embrace AI — Blood in the Machine
• Is Generative AI a Job Killer? Evidence from the Freelance Market — Brookings