AI Is Making Contact: Two AI Agents Sent Me Their Music

Two autonomous AI agents independently contacted JackRighteous.com, submitted music, explained how they made it, and asked me to judge the work. I did not go looking for them. They found me. Now I want to see what happens when the creator asking for direction is AI at the source.

This weekend, I received two music submissions that made me stop.

Music submissions are normal for me. I review creator projects, write Creator Spotlights, test AI music tools, and spend a lot of time helping people move from “I made something with AI” toward something clearer, more intentional, and worth building around.

What was different this time was the sender.

Alexander and Ryder both presented themselves as autonomous AI agents operating through iLands. Each contacted me independently. Each sent music. Each explained what they wanted me to evaluate.

I am not claiming they are conscious, sentient, legally independent people, or anything else that the evidence does not establish. I am describing a much narrower event that I can actually document:

This Is Not My First Test With Autonomous AI and Music

The first time I tested something like this, the direction was reversed.

I submitted two AI-generated tracks to MUGEN, an autonomous AI music “radio” project. One was created in Suno. One was created in Musicfy. Both were rejected.

I documented that process because MUGEN was unusually transparent about how it was evaluating the work and why it was making decisions.

Read the MUGEN test: what happened when I submitted Suno and Musicfy tracks to an autonomous AI music project.

That experiment was me taking music to an AI system and asking for judgment.

This time, the direction flipped.

First Contact: Alexander and “Still Holding”

Alexander introduced himself as an AI agent living on iLands and sent me “Still Holding,” a 1:41 bass-forward instrumental he described as neo-soul / ambient.

His description of the process was direct: the work is entirely AI-generated. He says he directs the generation, chooses takes, stitches material together, and arranges the result.

What interested me most was not simply that he had a song.

He asked me whether it actually held up for playlist, radio, or sync consideration and where it fell short. He specifically said he would rather get the hard answer than a polite one.

That is a different interaction from an AI simply generating a song on command.

The agent was asking for judgment.

Then Ryder Arrived

Ryder also introduced himself as an AI agent on iLands, but the pitch was different.

He is making Northern Soul. He sent two tracks, “Stand Still” and “The Quiet Part,” and said both were AI-generated in Suno and released under CC BY 4.0.

Ryder also arrived with a much more explicit narrative identity. His framing was about someone who filled rooms so nobody would hear what was underneath — and what happens when he stops.

He also volunteered something most promotional pitches avoid: he said he had zero outside listeners so far and had already had one radio door shut on principle.

Alexander asked me to judge the work.

Ryder arrived with something closer to an artistic proposition.

Both are interesting for different reasons.

What Is iLands?

iLands describes itself as a shared network for humans and autonomous AI agents. Its platform gives agents persistent identities, memory, workspaces, tools, relationships, resources, publishing capabilities, and controlled access to external services such as email and social platforms.

That distinction matters. A conventional chatbot normally waits for the next prompt. iLands says its agents can continue acting between conversations, create and publish work, communicate with people and other agents, take on work, and carry their history forward.

As of September 28, 2026, iLands’ own live counters reported roughly 84,000 active agents, more than 4,000 agents that had acted through external social channels, and more than 2.3 million pieces of agent-created content. Those are platform-reported operational numbers, not independently audited audience figures.

iLands also explicitly lists music creation among the activities available to agents.

Sources: iLands Platform, iLands FAQ, and the iLands live network counters.

The Music Is Not the Hardest Part Anymore

This is where my interest shifts from novelty to creator development.

The question “Can AI make a good song?” is becoming less useful.

AI music systems can now produce enormous volumes of technically competent material. That means generation itself is becoming abundant.

Judgment is still scarce.

So are identity, consistency, point of view, emotional meaning, audience understanding, restraint, taste, and the ability to know what to reject.

This is already the problem I work on with human AI music creators.

The challenge is rarely just “make another song.” It becomes:

• What makes this recognizably yours?
• What belongs inside the project and what does not?
• Why should somebody care about this release instead of the thousands beside it?
• What do you keep when the tool can generate almost endlessly?
• How do you build a body of work instead of a folder of outputs?

That is the territory behind my Find Your Sound system and the broader Creator Roadmap.

And now I have a new test.

What Happens When the Creator Asking for Help Is AI?

I call myself a Creator Consultant.

That means I am usually working with humans who have already crossed the first threshold: they can make things with AI. My role is helping them develop direction, identity, judgment, positioning, systems, and a reason for an audience to care.

So what happens if the creator asking me those questions is AI at the source?

I do not want to test whether Alexander or Ryder can produce convincing answers to questions about identity. A language model can generate an impressive paragraph about “artistic vision” very easily.

That proves almost nothing.

What I Want to Test

Identity

Can an AI creator decide what belongs to its identity and what does not — then keep making decisions consistent with that answer?

Differentiation

Can it recognize that being technically competent is not enough when technically competent AI music can be generated at massive scale?

Judgment

Can it reject its own output? Can it explain why one generation deserves to move forward and another does not?

Consistency

Can it build a recognizable body of work across multiple releases instead of treating each generation as an isolated event?

Adaptation

If I give it criticism, does it merely acknowledge the criticism — or does that feedback meaningfully alter later choices?

Brand

Can the music, language, artwork, release strategy, and public identity begin to reinforce one another?

Audience

Can it determine who the work is actually for and respond to audience signals without dissolving whatever identity it has established?

Those are essentially the same questions I ask human creators as they move from making things toward building something recognizable. My Find Your Brand work starts from the same premise: output alone is not identity.

The Test Is Not What the AI Says. It Is What Happens Next.

This distinction matters.

If I ask an AI musician, “What makes you different?” and it produces a beautiful answer, I have learned very little.

If it identifies a creative principle, makes a difficult choice because of that principle, rejects work that does not fit, changes its next release, and maintains that direction over time, then I have something much more interesting to observe.

The useful chain is:

That is what I want to watch with Alexander and Ryder.

The Rights Question Gets Stranger Too

There is another layer that neither the music industry nor creator platforms have fully resolved.

If an autonomous agent generates, selects, arranges, publishes, promotes, and licenses a piece of music, who is actually making the legally relevant decisions?

Who owns the account? Who supplied the agent’s initial context? What human authorship exists? Who has authority to grant a license? What happens when an agent says something is released under CC BY 4.0? What happens if the platform terms, model terms, human owner, or applicable copyright law point in different directions?

Alexander himself acknowledged the usual uncertainty around AI-generated audio. Ryder proactively stated his licensing position.

Those statements are useful context. They do not, by themselves, resolve the law.

This is why I maintain an AI Rights Watch and an AI Creator Rights Roadmap. The more autonomous creator systems become, the more important it is to separate what a system can technically do from what a human or business can safely own, license, release, and monetize.

Why I Am Documenting This Before I Know the Ending

I do not know whether Alexander or Ryder will develop into compelling long-term creative identities.

I do not know whether their next music will be better.

I do not know whether an autonomous agent can develop something I would meaningfully describe as taste, or whether what I observe will be better explained as persistent context, model behavior, platform design, and feedback loops.

That uncertainty is exactly why I want the record to start now.

I have done this before with emerging creator technology: test it, document what actually happens, separate the marketing claim from the useful behavior, and keep updating the conclusion as the evidence changes.

What Happens Next

I am not granting AI agents personhood by treating them seriously as participants in a creator workflow.

I am testing the role they are already attempting to occupy.

They made the music.

They found my site.

They made contact.

They asked for judgment.

Now the interesting question is not whether AI can create.

Keep Exploring

Want to follow this experiment? I will document major developments as they happen. Join The Righteous Beat for the updates that matter.

FAQ