Can a shortform show help mould public opinion around AI music?
One Minute Musicals: a show to humanize AI music
People hate AI music because it feels like a replacement to human artists
My loyal Attention Heads know that I historically have not watched movies, but that’s no longer a cute character quirk at my big ol’ age. I’ve been marching through the classics of film, and it brings me so much joy to ironically say things like, “ah, that reminds me of a line from Citizen Kane…”
But I am indeed now reminded of a line from Citizen Kane.
When the titular character Kane, a newspaper magnate, is pressed by his wife on what “the people will think,” he replies that people will think “what I tell them to think.”
The fact that the opinions of the many are controlled by the few isn’t all that interesting in and of itself. One of our great modern ironies is that we reject legacy media for its ability to selfishly sway public opinion, only to embrace individuals who, well, selfishly sway public opinion.
Lately, it feels like there are few bigger battlefields in the court of public opinion than AI. Now that the deepest pockets in the world seem to have won the perception battle in favour of LLMs, the opinion overlords are now turning their eyes to generative AI.
We shall think what they want us to think.
Meanwhile, twenty million people just digitally curbstomped an employee at Suno, an AI music company, for an admittedly tone deaf post about how money made music inaccessible to her as a kid, and how generative AI will soon change that. There are obviously real concerns around copyright issues in AI music, and I’m hopeful that some recent legal decisions will set a new precedent to pay artists for infringements, but admittedly, my apprehension of generative art is less a moral stance than it is one of taste.1
I suspect most people don’t hate AI art because of their morals; most people hate AI art because it’s dogshit. The lack of friction in its creation enables people with no “finger feel” for a real craft to mass-produce a bunch of slop. But that doesn’t have to be the case!
Take the song “I Run,” for example. It’s no Beethoven, but it’s a genuinely catchy tune that was unanimously loved on TikTok last month. Recently, some internet sleuths uncovered that the song is at the very least AI-assisted, and suddenly it’s divisive: some people refuse to listen to the song for moral reasons, while others are defending a song they enjoy irrespective of the process by which it was created.
Produce a shortform show that positions AI music as one tool among others in an artist’s toolkit
All of this to say: morality aside, AI music has a branding problem. The “I Run” debacle proves that there are enough people who will enjoy generative art as long as it’s entertaining and presented in an interesting context.2 I’ve been wondering if a company like Suno could accomplish this sort of massaging of the public opinion with a good shortform show. Let’s riff on one concept:
Title: One Minute Musicals (@oneminutemusicals)
Logline: We celebrate your most embarrassing stories with micro-musicals.
Format: A shortform scripted series. Every episode is a high production, self-contained story, similar to Submit Your Ick or FML Tales.
The production team solicits stories. Eventually it would be great to have fan submitted ones.
Suno hires a team of local playwrights and musicians to script a one minute musical based on the story. Importantly: Suno is one of the tools they use in the production process. They don’t use one-shot prompts to create a song, and instead use Studio mode, which is more comparable to a classic DAW.
Suno hires local talent to direct and perform the song/story.
It’s edited down into a one minute musical.
Casting: It’s important that every episode is produced and starred in by local artists. I imagine it’ll be a contentious topic, and many artists won’t touch it at all, but paying and highlighting real local creatives will at the very least be a gift to an audience that is very clearly hurt by generative AI. The key is likely to find artists who are already using generative tools in their workflows, and I’d imagine every episode would have credits listed in the caption, similar to Bilt’s Roomies.
Aesthetics: Something high-low would be fun here. I’m imagining a gothic, almost noir, visual aesthetic, contrasted against campy music over hyper contemporary stories. Epic Rap Battles of History would be on the moodboard.
I don’t think you can really know if a concept will land until you get some contact with reality. But there are a few explicit choices in the show’s design that are worth calling out:
I. Enable real artists to make something good
The win criteria for the show is simple but not easy. Suno needs to show that artists can create genuinely enjoyable things using generative AI as one tool in their toolkit. By including Suno in the process of creating a broader artifact, especially one that requires creatives across disciplines like acting and writing, you’d be able to subtly position the tool as the equivalent of a DAW or synth. A tool in service of the final vision, not a replacement to the process.
Some of the most popular shortform music shows, like Track Star or Ari At Home or Brita resonate because they explicitly highlight how music can make you feel: nostalgic, silly, goofy, communal. If Suno is going to change how people think about AI music, it needs to gift people a show that makes them feel something.
II. Designing a predictable payoff into the show
Earlier this week I compared Submit Your Ick to Bowlmates, two shortform dating shows designed with two very different approaches. One of the major differences, and subsequently one of the reasons I enjoy the former so much more than the latter, is because there’s a clear punchline you can expect in every episode. The “ick” getting revealed is a predictably interesting hook to come for, in the same way that people come to Subway Takes to hear a hot take or Boy Room to see a disgusting bedroom. Adam Faze, one of the early experimenters of “TikTok as TV,” says (@ 9:38) that “You’re watching because the hook is drawing you in, you have an incredible host that is keeping your attention, and there is a point at the very end that keeps them watching until the last frame.”
A mini musical on its own, if done well, might be entertaining enough to hook viewers and keep them around. But framing each episode around an “embarrassing story” is a final frame to keep people coming back for a predictable payoff.
III. A lens through which people interpret the world
The highest honour for a brand or company is to be associated with a product or action — you blow your nose with a Kleenex or you Google something. For a shortform piece of media, the equivalent is something like “becoming a new unit of storytelling.” AITA on Reddit, or FML, or, more recently, a Subway Take, all become new tools for you to interpret and package your own stories. Your sufficiently online friends don’t ask what your hot take would be, they ask what your Subway Take would be.
This is not to say a “One Minute Musical” is on the same tier as a “unit of storytelling” as AITA or FML or a Subway Take. But it’s important to intentionally design a show to become a lens through which people might interpret their own worlds. If Suno really wanted to win the coinage, they could probably run a botfarm that goes to any “Storytime” video on TikTok above 100K views and autocomments something like “lmao this should be on @oneminutemusicals.” But I do not think automations are winning any bonus points for an AI music company :)
In that same interview with Colin and Samir, Adam Faze talks about a show he launched called Bodega Run, a shortform gameshow in a bodega (@32:30). The first episode got 40M views across platforms, but subsequent episodes slowly dwindled down to the 50-200k mark. They realized that the show was a “stage for virality” if you had the right character, but the “format itself wasn’t that interesting for people.” Within 4 weeks, 10 episodes, and $3500, they decided to shut it down to pursue other projects.
It’s a good reminder that shortform shows are, as far as brand marketing goes, a low-risk bet. A higher production project like One Minute Musical would certainly cost more than $3500, but Suno could theoretically produce 3-6 pilot episodes, launch them on a brand new social account, and fold the experiment without any real consequences if the algorithm gods don’t pick up the video.
Thanks for reading! I explore how brands and startups build better media properties.
This post is part of my weekly strategy series, which is typically reserved for paid subscribers. If you enjoyed this, please consider upgrading to a paid subscription or sharing it with a homie.
Emotions run high in any conversation around generative AI, and I certainly don’t have all the answers, but I’m skeptical of anyone who claims that this isn’t even worth talking about. Let’s say that we hypothetically can’t stop the march of these companies training on public data. I wonder what the best case scenario would be for artists. The worst case scenario is obviously a world where all art is generated with no craft — everything would be either convenient background noise or hyper produced stadium hits. But is there a world where we can both use generative AI tools and create a world where more artists get paid for their craft. What would that look like? Is that paying artists based on their work’s presence in training data? Or some usage based model? I have no idea how that works, but it’s a worthwhile conversation that I am unfortunately not yet equipped to have. Pitching this shortform show isn’t an endorsement of how generative AI companies are treating artists; it’s one exploration of how they might start treating them better.
I fooled around with some generative AI tools last year, and had a few videos go pretty viral. I think the art direction was pretty good! It was surprising that, even with clear tells that the videos used AI for the imagery (jumpy visuals, some artefacts,) I bet something like less than 1% of comments even questioned how the videos were made, I got tens of thousands of comments and DM’s telling me how much they loved the work, and almost no one actively hated on the videos for using AI. These videos were pretty formative in my understanding of how people might embrace generative AI, as long as you’re creating something they enjoy.
I Run was such a great dress rehearsal because it actually didn’t use AI in the “computer, write me a tune” sense, but people discoursed as if