The real question was not “Can AI imitate an artist?”

The more useful question is: Can you use a recognizable artist as a temporary creative reference, learn from the craft, and then deliberately move the work toward your own expression?

That is the question I wanted to pressure-test with Melody Genie.

Artist references are tempting because they compress a lot of information. Say a famous rapper’s name and you may be trying to communicate rhyme density, aggression, internal rhyme, cadence changes, humor, storytelling, vocal attitude and production expectations all at once.

But that shortcut can also become a creative trap. If the generated lyric starts borrowing the artist’s name, biography, alter ego, recognizable story or other identity-specific material, you are no longer merely studying craft. You are carrying someone else’s identity into your song.

The artist reference should eventually become disposable. If the song falls apart when you remove the famous name, you probably have not developed your own direction far enough.

First: style, influence and copyright are not the same thing

There is no magic prompt, rewrite percentage or number of changed words that guarantees a song is legally safe. This article is educational and documents my creative process; it is not legal advice.

For U.S. copyright purposes, copyright protects original expression, while ideas, methods, concepts and principles themselves are not protected in the same way. A musical work and a particular sound recording are also distinct protected works. The U.S. Copyright Office makes both distinctions clear in its current guidance.

If you are working commercially, questions can extend beyond copyright into contracts, platform rules, publicity/identity rights and other law. When the stakes are meaningful, get qualified legal advice for your jurisdiction and your actual release.

Official background: U.S. Copyright Act §102 · Copyright Office guidance for musicians · Copyright and Artificial Intelligence.

For the broader JR breakdown, see my AI Music Rights & Ownership Guide and How to Use a Known Song as an AI Music Reference.

The starting brief

I intentionally did not begin with a finished Jack Righteous song. I gave Melody Genie a relatively simple concept:

For the first lyric pass, I used a famous rap artist as the Style DNA reference inside Melody Genie. That was deliberate: I wanted to see whether the system would stay at the level of technique or drift into identity.

It drifted.

The generated lyric included direct references to Marshall Mathers, Slim Shady, Dr. Dre and Detroit biography. That gave me a very clear line to work against. The first version was no longer just “dense New York-style rap writing.” Parts of the narrator belonged to somebody else.

Before Suno, I supplied the musical direction separately:

Notice what is missing from that Suno prompt: the artist’s name. I wanted to see whether the writing and craft direction could carry the idea without leaning on the celebrity label.

Hear the progression: three generations, three different jobs

Use headphones if possible and listen in order. This is not a contest between random generations. Each version represents a different stage of creator control.

What Melody Genie changed in the workflow

My earlier initial Melody Genie review asked whether the app could become more than another lyric generator. This dry run gives me the first real answer.

The first-generation button is not the most interesting part.

The more promising part is the revision environment. In the Studio Coach workflow I found controls for Weak Lines, Complex Wordplay, Multi Rhymes, Internal Rhymes, Vocalic Rhyme, Pattern Breaks, Specific Imagery, Hook Craft, Deeper Emotion and Story Arc.

I still have more to test. I have not yet found a clearly presented Suno-ready Style of Music prompt generated alongside the lyrics, so for this experiment I created the Suno production prompt separately.

The JR framework I would use with artist references

If you want a structured way to define the song before generating, use my Song Brief guide. If you already have a lyric and want to push it toward a finished project, use the Righteous Track Builder: Start From Lyrics.

There is also an AI-authorship question

The current U.S. Copyright Office position is that generative-AI output can be protected only where there is sufficient human authorship in the work; merely providing prompts is not enough by itself. Human-authored expressive elements, creative selection or arrangement, and creative modifications can matter. AI assistance does not automatically disqualify an otherwise human-authored work.

That is another reason I care about documenting the path from Version 1 to Version 3. The useful creator workflow is not “press Generate and pretend AI was not involved.” It is to make the human decisions visible: select, reject, rewrite, restructure, direct, compare and decide what survives.

Current U.S. Copyright Office AI guidance: Part 2 summary on copyrightability of generative-AI outputs.

My first takeaway from Melody Genie

I started this test asking whether Melody Genie could write usable lyrics before Suno.

I ended it with a more interesting question:

Can Melody Genie help a creator move through influence instead of getting trapped inside it?

Dry Run #1 suggests there is something worth continuing to test. The initial output crossed into another artist’s identity. The revision tools then gave me useful ways to attack weak writing and push the lyric farther away from that starting point.

That is not the same as saying the app makes a song original for you. The creator still has to make that happen.

My next tests will look at other genres, different lyrical voices, hook development, storytelling, how much manual editing remains necessary and whether the tool can improve a song that already begins with a strong creator identity.

Create What You Love | Love What You Create.
— Jack Righteous