Concept to Song: YouthPDX AI Music Workshop

The Challenge: Create An Educational Song

YouthPDX needed a session that could work for a group of kids with wildly different skill levels and zero assumed prep — no accounts, no devices to configure, no prior AI experience. It also had to produce something real: not a demo of a tool, but a finished, original song the group could point to and say "we made that."

The constraint that shaped everything: a single 60-minute window, live, with no second session to fall back on if something didn't render in time.

The Solution: Vote-Driven, Demo-Led Facilitation

I adapted iTK's existing "Learning Through Music Framework" — a self-directed, 5-step process for turning a technical topic into a memory-optimized song — into a live group format where kids participate by voting, not by operating tools themselves.

Core Design Decisions

• Pre-extracted concept options: I built out the full learning structure — core concepts, vocabulary, definitions, and what must be memorized exactly — for all 3 topic choices (Personal Branding, Growth Mindset, Financial Literacy) ahead of time, so nothing is improvised live once the group votes.
• Show the payoff first: A pre-built opener song plays in the first 18 minutes, so the group knows exactly what they're building toward before the live generation even starts.
• Vote, don't type: Kids shape the topic and the sound — genre, tempo, vocal style — through live Zoom chat votes, while I run the LLM and Suno steps on screen share, narrating every step in plain language.
• A backup that protects the ending: Suno's render time isn't fully predictable, so the opener/demo track doubles as the fallback for the final playback moment if the live track isn't ready.

The Pacing Map

Every minute of the hour is accounted for: an 8-minute framing intro on why music helps memory, a 10-minute demo, a 5-minute concept vote, 5 minutes of style votes (genre, tempo, vocal style), an 18-minute live-generation window where I run the song-structure, lyrics, and style-prompt steps live, 2 minutes to submit to Suno, and a closing playback and wrap-up.

The through-line I teach the whole way is the framework's own model: Music is the storage mechanism, repetition is the indexing system, lyrics are the knowledge, and melody is the retrieval cue — introduced in plain terms, then proven live on a topic the group picked themselves.

The Results

The session is built to hand a group of kids something they couldn't get from a lecture: proof, in real time, that turning a topic into a song is something they can watch happen and walk away with. No devices, no accounts, no prior AI experience needed — just a vote, a live build, and a finished track dropped in Zoom chat or sent to program staff afterward.

Instructional Design Principles

• Demo-led, not hands-on for tool operation: kids participate by choosing, not by typing prompts
• Show the payoff before the process: the finished example plays before the live build starts
• Kid-legible narration: every AI step gets a plain-language voiceover — no silent screen shares
• Always have a backup: the opener track protects the final playback moment no matter what Suno's render time does