OZ SONGS for Lyrical Literacy: AI Music Storytelling for Educational Impact
Professor Nik Bear Brown launches OZ Songs, a daily video journal reworking the Wizard of Oz with AI plus human craft, feeding techniques back into Lyrical Literacy.
Most tutorials about AI tools promise a shortcut: type a prompt, get a result, move on. Professor Nik Bear Brown opens the OZ Songs project with a different premise entirely. This is a public, ongoing journal of building a personal reimagining of the Wizard of Oz with AI and human craft together, documented one lesson at a time, precisely because the tools change so fast that a single polished tutorial goes stale within months.
A daily journal, not a finished course
The plan is not one big tutorial but a running series: short videos, three to five minutes, each teaching one specific thing as it comes up in the actual work. That might mean showing how to use a particular design tool, how to prompt effectively, how to build a custom GPT, or how to clone a voice, whatever the OZ Songs project happens to require that day. Rather than mapping out a fixed curriculum in advance, the series is structured as a working journal: build something, then make a video showing how it was done. Because Professor Brown teaches roughly five AI-related classes a semester at Northeastern, often involving games and art alongside more technical material, these same videos double as teaching resources for his students, who ask the same practical questions, how do I clone my voice, how do I composite an image, how do I write an agent, that this series is built to answer.
Building a consistent character
A recurring thread in the demonstration is Dorothy, redesigned as an original character rather than a direct copy of the classic book or film. The process is shown as genuinely iterative: an early version of her boots got swapped out for a pair of red Timberland boots because that better fit the desired look, and the same character was then placed in a poppy field background, with the point being that once a character design is locked in, it can be dropped into new scenes and settings while staying recognizably the same character. The video is explicit that this is a personal creative reinterpretation, not a faithful retelling. In the original book, Dorothy wears silver shoes and is roughly fifteen years old; the famous ruby slippers came from the film adaptation. This version keeps the red footwear but as Timberland boots, and ages the character up to around thirty, changes made deliberately in the spirit of making it a personal version of the story. Because the original Wizard of Oz material is in the public domain, adapting it this freely is part of a long tradition, one that has already produced numerous other reinterpretations of the same source material.
AI as a craft, not a shortcut
The video draws a direct comparison between AI work and writing: almost anyone can pick up a pencil and start a poem or a novel, but that alone does not make someone a working novelist or poet. Real novelists, poets, and songwriters refine their work repeatedly and care about the details, and that same discipline is what separates a quick AI output from something genuinely good. Typing a single prompt and accepting the first result will usually produce something merely okay. Treating AI generation like a craft, iterating, refining, and caring about the specific choices being made, is what the series is actually trying to teach, more than any single tool's interface.
Voice cloning with real stakes
Among the specific techniques mentioned, voice cloning carries the most personal weight. Professor Brown describes having cloned his own voice repeatedly for this kind of work, and mentions a striking example: taking an old recording of his father and using it to create a voice model so his father could speak again. It is offered as a concrete illustration of what these tools make possible, not as an abstract feature demo, and it is the kind of application that connects directly to why a project like this matters beyond entertainment value.
Who the series is for
The tone here is unusually direct about audience. The channel is explicitly for people with a positive, curious attitude who want to learn by doing; viewers who show up mainly to complain are told plainly this is not the channel for them. Professor Brown notes he does not read comments extensively but will act on ones that cross into pure negativity, while genuine questions get real answers. He also describes his background, a PhD in computer science, an MBA, multiple master's degrees, and work at think tanks starting at nineteen, as the reason his approach to teaching AI differs from the many channels that walk through button-by-button interface tutorials; this series is built around process and craft rather than a click-by-click guide.
Key takeaways
- OZ Songs is a running video journal, not a fixed course, documenting the daily process of building an AI-assisted reimagining of the Wizard of Oz.
- Short, focused videos teach one specific technique at a time, tools, prompting, custom GPTs, voice cloning, as they come up in the actual work.
- Character consistency is shown as an iterative process: a design choice like Dorothy's boots can be revised and then reused across new scenes and backgrounds.
- The Wizard of Oz is public domain, which is part of why this kind of free creative reinterpretation is possible.
- AI generation is framed explicitly as a craft, comparable to writing or songwriting, where refinement separates good output from a merely acceptable first draft.
- Voice cloning is demonstrated with a personal example: recreating a father's voice from an old recording so he could speak again.
Who this is for
This series is aimed at Professor Brown's own students, teachers, parents, and content creators who want to learn AI tools by watching a real, ongoing project get built rather than following a generic step-by-step guide. The techniques demonstrated, character design, voice cloning, and AI music workflows, directly feed into Humanitarians AI's Lyrical Literacy initiative, which uses music and storytelling to support literacy and cognitive development in learners of all ages.
Full transcript(auto-generated, with timestamps)
[0:01]Okay, Professor Bear here. I'm going to talk a little bit about this O songs project. Basically, I love, you know, uh, fairy tales and Wizard of Oz and stuff like that. And so, I'm basically going to make the whiz with AI plus humans. And I'm going to show my process every day. This is what this O uh sewing thing is about. So, what we're looking at here is just some design of the Dorothy character. So, I'll talk about how do I do this? How do I try to keep a consistent character? What works? What doesn't work? AI changes so quickly. That what might work in 3 months may not
[0:47]Work now. But what I do is I just do it. So, I'm a teaching professor for people who don't know at Nor Eastern. I teach five AA classes semesters typically. A lot of them involve games, art, things like that as well. Um, and so this will also be used for my classes. So when students ask, hey, how do I do this? How do I clone my voice? How do I, you know, composite? How do I whatever? Uh, how do I write an agent? How do I do this? How do I do that? Well, I'll have some videos for it. There's a lot of stuff. So my rather than try to do everything,
[1:21]My goal is to try to do a lesson a day. Think like a three minute, five minute video a day just teaching something. Teaching how to use Weebi, teaching how to use Hexfield, teaching how to prompt, teaching how to make a custom GPT, teaching how to clone your voice, whatever it is. Um, I'm not certain of what I'm going to teach, what what day, but this thing is basically like a journal, a video journal for me. uh which I think can help other people learn. One thing about learning is what I found about YouTube is kind of weird. There are a lot of people who go to a
[2:03]Free YouTube thing and just whine and complain. If that's you, don't come to my YouTube. Um if you have a positive attitude, great. We love you. Uh but I'm not going to do the bell thing, whatever subscribe thing. If you're one of these people who are just negative about everything, go to a different channel or a bajillion of them out there. uh don't go to my channel. Um and so if you are somebody who wants to learn, has a good attitude, then you're welcome. Uh but this is a lot for me and for students. And so this is a way of me just sort of
[2:37]Documenting my process because my big thing is learn by doing. And this is particularly true with AI because with AI the stuff changes so quickly that you know you just have to learn and learn and learn and learn and learn. Another thing about AI is a lot of people just think AI is writing a prompt and you're done. The analogy that I use is like poetry or writing a novel. Anybody can most anybody 99% of the world can take a pencil and start writing something. So they could start a poem and they can start a novel. But real novelist people are really good at it. And real poets people are really
[3:12]Good at it. Real songwriters people are really good at it. They refine and they refine and they're flying and they fine. and they care about what they do and they refine. That's AI. AI is a craft. Yes, you can, you know, get something. I'm going to stop this. I'll just set it going again. Or a little Dorothy thing here. It's I think poetry is a good analogy or writing is a good analogy. Anybody can sort of write something on a piece of paper. That doesn't make him Stephen King. And so there's a craft in the art to doing AI effectively. Yes, you can get something by just typing something in,
[3:49]But mostly what you'll get is just kind of okay. Uh, if you want to do something special with AI, you're going to have to treat it like a craft. You're going to have to treat it like a writer takes their craftsy or a poet or songwriter or anybody else. It's a air possibility. It's a very, very powerful tool. You can iterate so fast and do so things so fast and do things that just simply could have been done three or four years ago. like what I'm looking at now. This is a uh dworthy character that I designed uh to character I designed and now we're
[4:24]Doing things. So just go back a little scene here. So a little things I did is I didn't like her boot. So then I switched her boat to Timberland boots which I like. Um but we'll talk about how to do that. So if we look at these two things here, this is sort of the same picture except for I replace the boat and I put in a poppy field and I can put her anywhere. And what I'm going to do in this series is basically just do things. I'm going to do things for myself. That is I'm going to make this my version of the whist
[5:01]Or ricket or whatever you you know your favorite thing. Um, but I not document what I do. I'll make mistakes. I won't do everything perfect at AI will change. You know, I'm going to start this. My suspicion is I do a lesson a a day. I'm doing this the next year or two. There's a lot of stuff to land here. But the way to start is to start stuffer board. So, I'm just going to start, you know, as I do things that I need. For example, if I need to replace her boot. Yeah. I'll just make a video on how I do that. Uh, I need to create a voice for
[5:39]Dorothy. I'll show you how I do that. I may create a voice for Toto. I haven't thought about it yet. Uh, or just even barking noises, which is more likely, which it'll be actually easier to do uh, than have him speak. Um, but the fun thing is about you or whatever you want to do. My Dorothy. So, if you know the original Wizard of Oz, she actually has silver shoes, but the the famous people made it red ruby slippers, and I have red Timberlands, but in the real in the original book, she's about 15 years old. I made her about 30 years old just because I rather
[6:19]Fat. Um, I kept up. I changed, you know, a bunch of characters, but during the spirit of it, but it's my version. And you know that book was written I think in 1920 something maybe earlier. So it's you know well well into the public domain that's why there's so many things wicked in you know a bunch of things s Gregory or something read a bunch of books Disney you know everybody's you know this is the thunderous thing because it's just a beautiful series of books. So, I'm going to stop here. But basically, what this awesome thing is sort of the video jump. What I'm going to do is I'm going to
[7:00]Just do things and then record it and show people. I think people can learn from that, but I'm going to do it for myself. But if you look at my channels, and I have a three of them, one for my nonprofit, one for music, and one for my classes, and you like the work I do, then I'll show you how to do it. If you don't like the work I do, I don't know why you're on their channel. But if you don't like the work I do, there are people who come to my channel to just whine. Have a laugh. Get a life, dude. If
[7:34]That's what you do, do it somewhere else. If you have a positive energy, you're welcome to be at my channel. Uh I don't know if I can block people. I don't really read the comments that much. Um but if I can block you, I will. If you are irritating, uh, but if you were, you know, just a solid person or just don't say anything, that's great. But if you have questions, real questions, not just whining, then I'll try to answer them. Um, and mainly I'll show you what I do, my process, and the the way, you know, to do what I want to learn this guy's process. Look at the
[8:13]Work. If you like the work, I sculpt voices. I crepe these things. And if you hear it or you look at it and you like it, then you'll probably like, you know, the series was I'll show you how to do that even with your own voice, which I do quite a bit. um or somebody else's voice or your child's voice or I took a recording of my father and granted a voice from it so he could speak again. These are the things that you can do with AI and which is insane. It's insane what you can do with AI. So this is my sort of journal just sort of a public journal
[8:58]Of my making this you know my version of the kid or the whiz. Um, but because I'm a teacher primarily and a professor, I'm going to make it in such a way that I'll share my experience. And if you've ever been in one of my classes, and I teach five classes a semester and have probably 2,000 alumni students, um, it's all you by doing. You learn by doing. And so this is what this is going to be about. It's going to be about learning by doing it. So hopefully they let you sense of what this a song is about. It's my personal project plus it's me sharing my process and again
[9:38]It's my process. So I'll share my process but if it's something you know if you don't like my work then why you don't like my process. If you like my work you like the stuff that I do then my process is probably interesting to you because it's different than I think most people use AI. I have a PhD in computer science. I have an MBA. I have several master's degree started working at think tanks when I was 19 years old. So the way I do things is not the way other people do things. Most channel won't be all the other channels. There's bazillion channels which sort of take you through
[10:16]How do I click this button, how do I click that button. That's not this channel. Um but you you'll get a sense of how I teach and and what I teach as I do this. But this is just the intro. So this awesome project is about learning by doing and me sort of documenting my process while you all write custom tools all right individual AI tools to do a very specific task and I'll show you how to do it. So that's something that you might be interested in. You'll probably be interested in this. If you know if you look at my work and don't like it
[10:50]Then there are bazillion other channels and go to them. Okay. Uh hopefully that gets you a sense of what this oelang thing is about. I will uh begin, you know, as best as I can to try to get a video or so out a day or every other day. Take care.
More videos
2:08Bridging the Pixel Gap in Browser Automation.
2:23How One Narrow Safety Rule Can Make an AI Less Safe Everywhere Else.
2:04Why splitting a chunk from its document makes it retrieve for the wrong question
4:20Three You Can Take Back. One You Can't.
2:21Why a 50-turn agent pays for the same screenshot 35 times unless it caches the pixels
1:53