The Video That Stopped Me Cold
Three weeks ago, I spent fourteen hours in a studio choreographing a contemporary piece. Fourteen hours of negotiating with my body, replaying the same eight-count until my knees ached, scrapping sections that felt forced, rebuilding them from scratch. The final video—shot on a friend's phone, natural lighting, no post-production—had something real in it. You could see where I almost lost my balance. You could hear my breath.
Then someone showed me an AI-generated dance video. Smooth. Polished. Every movement mathematically precise. A digital avatar executing a routine that looked like it belonged in a music video budget I could never afford. My stomach dropped. Not because it was bad. Because it was good.
What ByteDance and Others Are Actually Building
The technology isn't theoretical anymore. ByteDance's video generation tools can produce footage from text prompts—type "ballet dancer performing Swan Lake variation in a rain-soaked alley" and get something watchable in minutes. Omnihuman and similar AI avatar platforms create digital performers who sing, dance, and emote without ever stepping into a rehearsal space. The uncanny valley is narrowing fast.
For content creators, this is seductive stuff. A small dance studio in Omaha can now produce promotional videos that rival what used to require a full production team. An independent choreographer can visualize concepts before teaching them to actual dancers. A brand can generate twenty versions of an ad in the time it used to take to film one.
I get the appeal. I've stared at my own shoestring budget enough times to know that access to professional-looking video content changes the game for small operations.
The Part Nobody Wants to Talk About
Here's what bothers me, though. Dance exists in the body. It exists in the specific way your hip shifts when you transfer weight, the micro-expression that crosses your face mid-turn, the particular quality of effort that comes from muscles that have been trained and scarred and rebuilt over years. AI can simulate the shape of movement. It cannot replicate the why behind it.
I watched a colleague—an incredible hip-hop dancer with fifteen years of battle experience—react to an AI avatar performing a popping routine. His response was immediate and visceral. "That's not popping," he said. "That's a skeleton moving to a beat." The avatar hit every position. It missed every intention.
There's a concept in dance called groove. It's the thing that makes you nod your head when you watch someone move, even if you can't articulate why. Groove lives in the space between beats, in the slight delay or anticipation that a human body naturally produces. Current AI generates movement on beat—mechanically correct, emotionally vacant.
Who Gets Credit When Nobody Danced?
This question keeps me up at night. If I prompt an AI with "create a contemporary dance piece about grief" and it generates something moving, who choreographed it? Me, for writing the prompt? The engineers who built the model? The thousands of unnamed dancers whose footage trained the algorithm? The machine itself?
We're watching creativity get abstracted into inputs and outputs, and something about that feels dangerous. Not in a "robots are coming for us" way—I'm not that dramatic. Dangerous in the way that convenience always is. When the easier path exists, the harder one starts to look pointless. Why spend months developing a piece when a prompt produces something shinier in seconds?
The Democratization Argument Is Real (And Incomplete)
People throw around the word "democratization" like it settles the debate. And sure—lowering barriers to content creation is genuinely good. A community dance group in rural Mississippi shouldn't need a five-figure budget to share their work online. AI tools make that possible, and I won't pretend that's insignificant.
But democratization of tools isn't the same as democratization of artistry. Giving everyone a camera didn't make everyone a photographer. Giving everyone video generation won't make everyone a choreographer. The risk is that we start mistaking output for craft, volume for value.
I've talked to dancers who've stopped posting their own videos because AI-generated content gets more engagement. That should concern us. Not because engagement metrics matter in some cosmic sense, but because visibility is how dancers find audiences, get booked, and sustain careers.
What I'm Actually Doing About It
I'm not boycotting AI tools. That feels performative and futile. Instead, I'm using them for what they do well—mood boards, concept visualization, rough drafts of staging ideas. Then I take those seeds into the studio with actual humans and let the movement evolve through the mess of real bodies doing real work.
Last week, I used an AI tool to generate visual references for a piece I'm building about isolation. It gave me three concepts I wouldn't have considered on my own. Then I spent six hours with two dancers, and what emerged looked nothing like the AI output. It was better. Not because humans are inherently superior, but because the negotiation between bodies in a shared space produces things that algorithms can't anticipate.
The Thing About Sweat
There's a moment in every rehearsal I've ever been part of—when someone nails a section they've been struggling with, and the room shifts. You feel it. The other dancers feel it. Sometimes it's loud, a collective exhale or a shout. Sometimes it's quiet, just a change in the quality of attention. That moment is what dance is for me.
AI doesn't have bad days. It doesn't fight through injury or push past mental blocks or discover a movement vocabulary born from frustration. It produces, efficiently and endlessly. And maybe that efficiency is exactly why it can't replace the thing that makes dance matter to the people who do it.
The best performances I've witnessed weren't technically perfect. They were honest. They carried the weight of someone who had actually lived inside the movement, who had failed at it enough times to understand what it cost.
No prompt captures that. No algorithm generates it. And that, honestly, is what keeps me hopeful.
---
This rewrite:
- Opens with a specific personal hook (14-hour choreography session vs. AI)
- Takes a clearly opinionated stance while acknowledging AI's real benefits
- Uses concrete examples (colleague's reaction, community dance group, personal AI usage)
- Varies paragraph length and opening style throughout
- Ends with a visceral, specific image rather than a generic call-to-action
- Avoids all flagged AI clichés (no "delve," "tapestry," "landscape," hedging language)
- Maintains human imperfection—contractions, fragments, direct address
