■ SAFE RISK ■ Creative Arts
No. Motion-generation models can output endless movement phrases, but choreography is made on and with living dancers — coaxing material from their bodies, shaping it into meaning, and getting a company to perform it like they believe it.
“AI generates movement. But expressing human emotion through dance is still... human.”
Our AI replacement risk score — how we score jobs
Contemporary choreography happens in a studio, not a timeline. The working method for most choreographers is collaborative: setting improvisation tasks, watching dancers generate material, selecting and distorting phrases, structuring them against music and silence, and rehearsing until intention reads from the back row. Around that core sit the unglamorous duties — grant applications, commissioning negotiations with companies and festivals, lighting and costume collaboration, teaching company class, and restaging older works on new casts.
AI has entered the studio as a tool with a real lineage — Merce Cunningham was using chance procedures and computer software decades before it was fashionable. Today's motion-generation models produce movement sequences from text or music, and companies like Wayne McGregor's have trained AI on their own archives to suggest phrases in the house style. Notation, video analysis, and archival restaging also benefit. For a choreographer facing empty-studio dread, a machine that proposes strange material is a legitimate collaborator.
What can't be delegated is everything that makes the material a work. Movement generated on a screen must be translated onto specific bodies with specific injuries, ranges, and qualities — choreography is fitted like tailoring. The choreographer's authority in the room, the dramaturgical sense of when a phrase should break, the relationship with dancers that gets them to take risks — these are relational crafts. And the market is brutally reputational: commissions flow to names and visions, not to movement libraries. The economics were always the hard part; AI changes them barely at all.
Automatability: our editorial assessment of current and near-term AI capability
Resilient for the foreseeable future. AI movement generation is entering studios now as a creative tool — expect it to be a normal part of the process by 2030, the way video and notation software became. The choreographer's role directing dancers, shaping works, and holding artistic vision faces no displacement on any visible horizon; the profession's chronic threats remain funding cuts and touring economics, not models.
It can create movement — models generate dance phrases from music or text, and some are trained on specific choreographers' archives. But movement material is raw ingredient, not choreography. Turning phrases into a work requires fitting them to real dancers' bodies, structuring them into meaning, and rehearsing them into performance. Choreographers who use AI describe it as a strange, useful collaborator: good at suggestions, incapable of intention.
AI barely registers among the profession's actual risks — arts funding, touring costs, and the small number of salaried posts have always been the hard part. If anything, motion-AI adds work: tech companies, game studios, and virtual-performance projects hire choreographers and movement directors to make digital bodies move believably. The career remains difficult for the same old reasons, not new ones.
It's a live concern. Motion-capture datasets and performance videos feed movement-generation models, and dance's copyright protections are historically weak — choreographic works are protectable, but movement style isn't. Practical steps: control where full works are posted, register significant works, and negotiate data terms in commissioning contracts. Dance organizations are beginning to push for consent standards, but the law is well behind.
As a generative sparring partner. Wayne McGregor's company trained a model on decades of his archive to propose phrases during creation — he keeps, warps, or discards them. Others use AI for music editing, video analysis of rehearsals, and restaging from archives. The consistent pattern: AI expands the option space early in a process, and humans do all the choosing, shaping, and rehearsing thereafter.