What are some examples of AI-generated music?

The Chart-Toppers: Named AI Artists You Can Actually Listen To

The first AI-generated artist to hit #1 on a Billboard chart isn't a lab experiment — it's a country song. Breaking Rust's "Walk My Walk" topped Billboard's Country Digital Sales chart, marking what Ohio University's Josh Antonuccio called "an important indicator of how AI-music is starting to generate money and attention for the creators" (Ohio University reporting).

And Breaking Rust isn't alone. On one recent Billboard Country Song Sales Chart, one third of all songs were AI-generated — including the #1 and #3 tracks (according to the same report).

The Velvet Sundown tells an even stranger story. This AI-generated indie rock "band" went viral in June 2025, landing on popular Spotify playlists with no social media presence, no interviews, and no live performances. It amassed nearly a million monthly Spotify listeners and released two albums before being revealed as AI-generated (per Carnegie Mellon University).

Other named AI acts have crossed commercial thresholds that would make human artists envious:

• Xania Monet — an AI-generated R&B avatar that earned enough radio airplay to debut on a Billboard radio chart.
• Cain Walker — an AI country act topping digital sales charts, alongside Breaking Rust.
• Ennly Blue — an AI artist whose top song cracked 10 million listens.
• Solomon Ray — a fictional AI gospel singer built by rapper Christopher "Topher" Townsend using ChatGPT for lyrics and Suno for song generation, producing Billboard-chart-topping gospel music (as Courthouse News reported).

Townsend's pitch for his synthetic singer is blunt: "He has an immaculate voice. He doesn't get sick... and he can work faster than I can work." That's the commercial logic driving this wave — infinite output capacity at near-zero marginal cost.

The scale behind these chart appearances is staggering. Suno and Udio together generate over 10 songs per second, and 60 million people used AI software to create music in 2024 (per the IMS Business Report). Deezer reported more than 20,000 fully AI-generated tracks submitted daily — nearly double its figure from four months prior.

For anyone tracking this space, the pattern is familiar to us at SkillzBench: a flood of AI-generated output hitting distribution channels faster than anyone can verify what's real, what's derivative, and what's safe to trust. AI music is no longer a novelty — it's competing directly with human artists on streaming platforms, and the charts now prove it.

The Viral Fake: 'Heart on My Sleeve' and the Takedown Era

In April 2023, a track called "Heart on My Sleeve" — supposedly a Drake and Weeknd collaboration — exploded across social media, and it fooled a lot of people. The catch: neither artist had anything to do with it. An anonymous creator known as Ghostwriter977 used AI to clone both vocalists, and the song racked up 600,000 Spotify streams, 15 million TikTok views, and 275,000 YouTube views before Universal Music Group moved to pull it from TikTok, Spotify, and YouTube on April 17, 2023, as reported by The Guardian.

The fake duet wasn't an isolated stunt. AI-generated Drake vocals began appearing grafted onto other artists' hits, including Ice Spice's "Munch," Cardi B and Megan Thee Stallion's "WAP," and Bryson Tiller's "Don't" — a wave of unauthorized "ghost verses" that turned one of music's biggest stars into a involuntary collaborator. Drake's response was blunt: in an Instagram story reacting to the AI-assisted "Munch" remix, he declared it "the final straw AI," per the same Guardian reporting.

Why did "Heart on My Sleeve" get pulled while other AI songs went on to chart? The dividing line is largely about whose voice and whose rights are being used:

• Voice cloning of real, named artists triggers label takedowns on intellectual property grounds.
• Original AI acts like The Velvet Sundown — which drew nearly a million monthly Spotify listeners — faced no such claims, since no identifiable artist was impersonated.
• Licensing changed the game: Universal and Warner settled with Suno and Udio, turning some former defendants into licensed partners.
• Consumer pressure is building for disclosure — 81.5% of surveyed UK listeners say AI-generated music should be clearly labelled.

That takedown-versus-chart split reveals the real fault line: impersonation and unlicensed training data invite legal action, while original AI-generated acts operating under licensing deals are increasingly treated as legitimate competitors. As Ohio University's Josh Antonuccio noted, AI music is now "competing directly with human artists on streaming platforms" — an entirely different ecosystem from the 2023 novelty era.

For anyone building in this space, the lesson is that provenance matters as much as output. Just as SkillzBench is being built to verify which AI agent skills actually perform and which carry hidden risks, the music industry is developing "fingerprinting" systems to mark AI-generated content at the point of creation — because knowing what's real, and what's safely sourced, is becoming the whole game.

How These Songs Are Made: Suno, Udio, and the Text-Prompt Pipeline

Behind every AI-generated hit sits a surprisingly simple pipeline: type a prompt, get a song. Platforms like Suno and Udio let anyone generate complete, radio-ready tracks with no musical training — Udio famously produced an '80s-style rock anthem about monkeys on the moon from just a few words of text, according to researchers at Carnegie Mellon.

The scale is staggering. Together, Suno and Udio generate more than 10 songs per second, and an estimated 60 million people used AI software to create music in 2024. That output explains how acts like The Velvet Sundown and Breaking Rust materialized seemingly overnight — complete with albums, cover art, and streaming playlists.

The tools themselves are evolving fast. Suno released its first models built with licensed music from Warner Music Group, BMG, and Believe in September 2026, while Udio holds licensing deals with Universal and Warner, per the Los Angeles Times. The shift from training on scraped catalogs to licensed material marks a turning point in how these songs are legitimately made.

Not every workflow is fully automated. Assistive tools like Lemonaide handle lyric generation, mixing, mastering, and stem stripping, cutting song completion time from 6–12 months to just 2–4 months. In these setups, AI functions as a production accelerator rather than a replacement — closer to the verification-and-quality-judging mindset we apply to agent skills at SkillzBench, where the tool is tested against a human-defined standard rather than trusted blindly.

• Text-prompt generation: Suno and Udio turn a few descriptive words into complete songs with vocals, lyrics, and instrumentation.
• Licensed models: Suno's Warner, BMG, and Believe partnerships mean newer outputs are built on legally cleared training data.
• Assistive workflows: Tools like Lemonaide compress production timelines from months to weeks by automating mixing, mastering, and lyric drafting.
• Hybrid creation: CMU's Amuse transforms images, text, or audio into chord progressions for human-AI songwriting collaboration.

The pipeline matters because it shapes what listeners hear. As reporting on Suno and Udio notes, users have flooded the internet with millions of AI-generated songs, some of which land directly on Spotify alongside human-made tracks. Understanding how these songs are made — and which ones rest on licensed foundations — is quickly becoming essential context for anyone navigating modern streaming catalogs.

The Catch: Lawsuits, Licensing Deals, and the Creativity Gap

For every AI track topping a chart, there's a courtroom battle, a licensing negotiation, or a listener demanding to know whether a human actually wrote the song. The friction around AI-generated music is as much a part of the story as the music itself.

The trouble started in 2024, when Universal, Sony, and Warner all sued Suno and Udio over alleged training on copyrighted recordings. Since then, the industry has pivoted from litigation to dealmaking: Universal settled with Udio, Warner settled with both platforms, and Suno released its first models built with licensed music from Warner, BMG, and Believe, according to reporting on the licensing shift. Not everyone has come to the table, though.

Sony remains the holdout. The label filed a second lawsuit against Udio covering more than 30,000 recordings, and Sony and Universal filed a new suit alleging Suno's latest model still infringes. A Munich court also ruled against Suno in a GEMA case in July 2026. Meanwhile, the American Federation of Musicians sued Universal and Warner over licensing session musicians' performances without consent — a reminder that even "settled" deals leave open questions about who actually gets paid.

Artists themselves haven't been quiet, either:

• More than 1,000 British artists released the silent album "Is This What We Want?" in February 2025 to protest U.K. copyright law changes.
• The Artists Rights Alliance organized a "Stealing Isn't Innovation" campaign backed by artists including Cyndi Lauper and Bonnie Raitt.
• Composer Max Richter told UK MPs that training AI on copyrighted music is "unfair and unworkable" for artists.

Beyond the courtroom, there's a creativity question. A Carnegie Mellon University study had 140 musically trained participants create short melodies, with some using Udio for inspiration. The AI-assisted music took longer to make, used fewer notes, and was judged less creative than purely human-composed work. CMU's Richard Randall puts it bluntly: AI music "is always going to be derivative in some way, it's always going to be playing it safe."

Listeners seem to agree. A BPI/AudienceNet survey of 1,750 UK music consumers found that 81.5% believe AI-generated music should be clearly labelled, and over 80% value human-made music more than AI-generated music. AI companies are responding with "fingerprinting" systems designed to mark AI content at the point of generation.

That demand for transparency is familiar territory for SkillzBench, which is building verification and benchmarking infrastructure for AI tools on the premise that outputs should always be traceable and clearly identified. In music, as in software, the question isn't just "what can AI make?" — it's "can you prove what it made, and how?"

The emerging consensus positions AI as inspiration and ideation rather than full replacement. Tools like Lemonaide help artists cut song completion time from 6–12 months to 2–4 months, and CMU's Amuse transforms images or text into chord progressions for human-AI collaboration. The most durable AI music may be the kind where a human still holds the pen.

How to Tell AI Music from Human Music: Verification as the New Skill

When The Velvet Sundown racked up nearly a million monthly Spotify listeners, nobody knew the "band" didn't exist. The reveal only came after the act had already landed on popular playlists, released two albums, and quietly earned more than a million streams — no social media, no interviews, no live shows, no humans.

That's the world listeners now navigate. Platforms like Suno and Udio generate over 10 songs per second, and Deezer reports more than 20,000 fully AI-generated tracks arriving daily — nearly double the count from four months earlier. Ears alone are no longer a reliable detector, which is why verification is becoming a core skill for anyone consuming or deploying AI output.

The industry is responding on several fronts:

• AI companies are developing fingerprinting systems that mark AI-generated content at the point of generation, embedding provenance into the file itself.
• Consumers are demanding disclosure: in a BPI/AudienceNet survey of 1,750 UK music listeners, 81.5% said AI-generated music should be clearly labelled.
• Streaming platforms are building their own detection and labeling approaches, with Deezer already tagging AI submissions at scale.
• Licensing deals from Warner, Universal, and others now shape which AI tools operate legitimately — provenance is becoming a commercial requirement, not just an ethical one.

But labeling only works if someone checks that the label is true. The "Heart on My Sleeve" fake Drake track amassed 600,000 Spotify streams before platforms pulled it — detection arrived after the audience did. The broader lesson extends well beyond music: AI outputs, whether a song, a code snippet, or an agent skill, need independent verification rather than trust by default.

That principle drives how we're building SkillzBench. Published research found roughly 26% of AI agent skills carry at least one vulnerability, yet most users install them blind. Our approach mirrors what the music industry is learning the hard way: pin every version, scan for security issues, benchmark performance in controlled conditions, and publish a trust score that money can't influence. Verification has to be systematic, not vibes-based.

For readers who want AI tooling they can actually trust — tested, scored, and transparently benchmarked — the SkillzBench waitlist is open. Verification as the new skill starts with tools that have already earned it.