I write the songs. AI helps me give them a voice.

I make music because I enjoy it. I’m not a career musician, I don’t earn much from it, and I’m not trying to become a superstar. I play instruments, write songs and have some audio-engineering and remastering experience. AI has helped me produce and release my songs, including vocals I could not perform the way I wanted. For this song, I used an AI clone of my own voice: the voice is based on mine, but I did not physically sing the performance.

Mostly, I’m having a lot of fun with it. If someone listens to one of my songs, enjoys it and feels entertained, that is a good feeling. A song that might otherwise have stayed an idea can become something somebody hears. That is why I see something worth celebrating in this technology.

A song about enjoying your own company

I’m a single bloke in my 50s who enjoys his own company. Friends don’t drop in or call as often as they once did; much of our contact happens through social media. Nobody is to blame. People have families, work and busy lives.

I’m not depressed or lonely, but boredom creeps in. I take on random projects simply to have something to do. Another creative project? Why not. Digging a hole in the backyard just because I can? That is the sort of thought that inspired this song.

I suspected I wasn’t the only bloke my age who recognised that feeling. There is comfort, and a bit of humour, in knowing other people are in the same boat. I wanted something relatable and entertaining, rather than a sad song about getting older.

I knew the genre, instruments and overall sound I wanted: alternative dance, bass-driven music and house are big influences. Getting AI to deliver what you hear in your head takes specific direction and trial and error. Understanding instruments, arrangements and how sounds fit together helps.

They are my words, my voice and my choices about the sound and delivery. AI helped bring those ideas together into something I could share. I have two other versions I enjoy too; the excerpt here is the Deeper Than Planned Remix, not a before-and-after demonstration.

In my Apple Music artist view, this is currently my top song within my own catalogue. That is a personal observation, not a claim about the public charts. More than a ranking, I enjoy the thought that something about it is connecting with someone.

Some people will reject AI music whatever it sounds like. This article is for the people willing to listen, and for musicians wondering what they can now make. There is real progress here. There are also new rules about how that music reaches an audience, and understanding them gives creators more control.

“I don’t want to be a superstar. If someone listens to one of my songs, enjoys it and is entertained, then that’s a good feeling.”

The bigger picture

Music has been using digital help for decades

Digital remastering, restoration and production tools have long helped artists achieve results they could not easily produce unaided. Abbey Road traces its work with CEDAR restoration technology to the late 1980s. More recently, iZotope dates Ozone’s AI-powered Master Assistant to 2017.

Those tools do different jobs from generating a lead vocal. Earlier digital processing was not all AI, and polishing a performance differs from creating one. But the longer history matters: musicians using technology to extend their capabilities is thoroughly familiar. Today’s tools make more of the creative process accessible to more people.

“AI music” covers very different creative choices

Writing lyrics, directing an arrangement, generating a vocal and mastering a recording are different jobs. Calling all of them “AI music” conceals the decisions that went into the song.

The three categories below are used in ARIA’s current guidance. They are useful distinctions, but they are not a universal scoring system for every platform or a verdict on whether a song is enjoyable.

ARIA excludes the first category from its charts; the other two can qualify subject to its remaining rules. The change applies from the chart dated 31 August 2026. Chart eligibility does not decide distribution, radio play or royalties.

Separately, the 10 July voluntary industry labelling proposal distinguishes AI-generated from AI-assisted recordings. It does not cover AI use in lyrics, composition, videos or cover art. Human-written lyrics with a generated lead vocal can therefore still fall into its AI-generated recording category.

That distinction matters to me. The label can describe how the vocal was made without telling the whole story of the songwriter behind it. I want accurate credits that leave room for that human contribution.

The industry is building around AI

Record companies are doing more than debating the technology. Warner Music Group and Suno’s November 2025 agreement settled their litigation and set out plans for licensed models and participating artists’ control over uses of their identity, voice and work.

Universal’s 10 September agreement with ElevenLabs adds another route: a planned, separate platform for participating artists’ music, including remixes and new interpretations. Announcing an agreement is different from making every proposed experience available today.

AI acts have attracted label deals too: CBS’s 2025 interview with Xania Monet’s creator reported the project’s recording deal with Hallwood Media.

These developments are a meaningful commercial commitment. They do not mean every label shares one position, or that an independent creator is guaranteed a deal. My reading is that the industry increasingly sees something it can build a business around. The question for musicians is how much creative freedom, access and value they retain as that business develops.

Being available is different from being discovered

Getting a song onto a service is only part of reaching listeners. Credits, recommendations, search and artist identity now have their own AI policies.

There is a worthwhile distinction between giving listeners information and deciding on their behalf what deserves to be heard. I support honest disclosure. I would also like well-made music to have a fair chance of finding people who enjoy it.

Spam and fake listening are separate problems. Spotify’s 20 October update describes action against mass uploads, duplicates and manipulation, while acknowledging the limits of AI detection. Those behaviours deserve scrutiny. Using a vocal tool is not, by itself, evidence that a musician is running a content farm.

For everyday listening, there is also a wonderfully ordinary response to a song you do not enjoy: press next. A crowded catalogue can make discovery harder, but we already skip human-made music that does nothing for us. I would rather give a song a chance and move on when it misses than dismiss it before hearing it because of the tools involved.

Inspiration is not the same accusation as sampling

Musicians have always learned from what they hear. Influences, genres and shared musical traditions are part of making music. I understand why creators object when “you used AI” is treated as a complete argument that they stole something.

A specific accusation should identify what was copied. Sampling has a concrete meaning: reusing part of a recording. APRA AMCOS explains that clearing someone else’s sample involves permissions for the recording and underlying work. A familiar mood or a general genre comparison does not, on its own, demonstrate that a particular sample was taken.

Training permissions are another question, as are copied lyrics, a reproduced melody or an impersonated voice. The U.S. Copyright Office’s separate reports on training, output copyrightability and digital replicas illustrate why these issues should not be bundled together. They address U.S. law, not a universal rule.

The analogy with human inspiration deserves discussion; it does not automatically settle a training dispute. Equally, a dispute about a model’s training does not itself demonstrate that every person using it has pasted a stolen sample into their song. Be specific, show the evidence and distinguish allegations from findings.

Back to making music

Use the opportunity well

If you have songs waiting to be heard, there is a practical way forward. Start with the part you want help with: a vocal, an arrangement, production or mastering. Keep the choices that make the song yours, and be prepared to edit or discard a result that does not serve it.

1. Finish the song deliberately. Listen for the meaning of the lyric, phrasing, pronunciation and whether the performance suits it. A polished sound is not the same as a satisfying song.
2. Keep a simple creation record. Save lyric drafts, recordings, stems and edits. Note what you performed, what you generated and which version you released.
3. Check the release route before committing. Confirm the tool’s commercial-use terms, export access, distributor acceptance and requested disclosures. A subscription alone does not resolve every rights question.
4. Describe the contribution honestly. “Lyrics by me; generated lead vocal; arrangement and final edit by me” can be useful when true. Do not claim a human performance that never happened.
5. Build an audience around the music. Share the story of the song and invite people to listen. Judge the response by real engagement, not bought streams or a folder full of outputs.

Check exports particularly carefully. Udio’s current partnership help page, dated 17 February 2026, says audio, video and stem downloads were disabled. It is a reminder to verify present capabilities rather than assume an earlier workflow still exists.

“AI has helped turn the music in my head into something other people can hear.”

AI has helped me produce and release songs, and made the process a lot of fun. I want more people to have that opportunity. A song does not have to make its creator famous to be worth making. If it gives somebody a few enjoyable minutes, that matters too.

Sources, scope and disclosure

Research checked 21 October 2026. Primary policy and company announcements are linked beside the claims; the Xania Monet example is attributed to CBS’s interview reporting. Announced rollouts are distinguished from available features. No streaming account, detection system, music model or private contract was independently tested.

The author makes music for enjoyment, plays instruments, writes songs and has some audio-engineering and remastering experience. He earns little from it and is not pursuing stardom. He uses AI to help produce and release songs, including vocals. The song excerpt and account of its creation were supplied by the author. Its generated vocal uses a clone of his own voice; no live singing performance is claimed. The Apple Music observation concerns his own catalogue and was not independently verified. The artist name and profile links are intentionally omitted: this is a creative example, not an artist promotion. That perspective informs the opinion here; it is not evidence of market-wide results. AI tools assist our research and publishing workflow, with strict controls around human input, review and editorial approval.

The backyard image is an AI-generated editorial illustration inspired by the song, not the author’s property. The studio image used in article previews is also an AI-generated editorial illustration, not the author’s studio. Hand-drawn musical flourishes are decorative vector artwork, not notation or an audio waveform extracted from the song. Platform and chart rules can change; follow the linked current guidance for a particular release.