Animator · The artist who decides what happens between the drawings — from Reynaud's hand-painted Paris projections to a Latvian Oscar winner made in free software.
Animation is more exposed to automation than most creative professions, and pretending otherwise helps nobody entering it. The output is fully synthetic by definition — there is no camera, no location, no performer whose physical presence has to be arranged — so every part of the process that can be specified in advance is a candidate for a model, and several already have been.
The parts that resist are narrower than the profession would like but sharper than the panic suggests. A generative system can produce plausible motion; it cannot yet decide what a character has concluded one frame before it moves, hold that decision consistently across four hundred shots, or take a director's note and change it. Those three things are what a studio is actually buying.
Close to half of what animation departments do is mechanical execution downstream of a creative decision: in-betweening, clean-up, colouring, matting, rotoscoping, lip-sync, motion-capture cleanup and background element generation. Most of that is already automated or automating fast, and the junior rungs where it lived are disappearing first. The performance decision, the consistency of a character across a whole production, and the ability to respond to direction remain outside what current systems do reliably or can be held responsible for.
Scored from the tasks, not the job title. Lower is safer.
Jobs AI cannot take →Choosing what a character has concluded, and letting the audience see that conclusion arrive before the body responds, is a judgment about intent. Generative systems produce motion that is plausible without being about anything.
Studio work is iterative: a director says the hesitation is half a beat late and the animator changes exactly that. Systems that regenerate rather than adjust cannot participate in a note cycle, which is where most production time actually goes.
A character has to move the same way in shot 12 and shot 380, eighteen months apart, across several animators and often several countries. Holding a performance standard over that span is a supervisory human function.
Stop motion, puppet animation, sand and paint on glass, cut-outs under a downshooter. Their whole value to an audience is that a person did it by hand, which no synthetic version can supply even if it looks identical.
Someone has to hold copyright, sign off delivery and answer for what is on screen. Training-data provenance and the legal status of machine-generated frames are unresolved in most jurisdictions, and studios are cautious accordingly.
Already gone in most pipelines. Digital ink and paint erased the department in 1990, automatic line closure and fill handle the rest, and machine matting now does rotoscoping work that used to occupy whole floors.
Vector in-betweening tools and, since the 2020s, frame-interpolation models generate the frames between an animator's poses convincingly for most non-deforming motion. This was the craft's apprenticeship rung, and it is closing.
Phoneme-driven mouth animation from an audio track has been solved to broadcast standard for years, and facial performance transfer from a recorded actor to a rig is now routine in games and episodic work.
Procedural and generated environments, crowd simulation, and automated cleanup of noisy motion-capture data absorb work that once required large teams — with quality still uneven enough to need a human pass.
Text-to-video output is not shippable as final animation, but since around 2024 it has been good enough for pitch reels, mood pieces and previsualization — which is where a meaningful share of paid short-form and advertising work used to sit.
Game engines now render broadcast-quality animation as it is played back, so the loop between a change and seeing it collapses from hours to seconds. That changes how animators work more than it changes what they decide.
Interactive and episodic work already employs far more animators than feature film, and demands a different skill set — cycles, transitions, state machines and gameplay feel rather than a single authored shot.
Following the 2023 writers' and actors' agreements in the United States, the Animation Guild's 2024–25 bargaining round put notification and consent language about generative systems into an animation contract for the first time. Other territories are watching the precedent.
Builds the puppets other animators perform with — skeletons, deformation, cloth and hair — and writes the tools around them. Scarce enough that it is often the best-paid non-supervisory role in a 3D studio.
Blocks whole sequences roughly in 3D so a director can edit a film before anything is built. Grew out of animation departments and now serves live-action and virtual production as much as animated features.
Game engine character work, plus the Live2D rigging and motion-capture puppetry behind virtual performers — a branch that barely existed before 2016 and now employs thousands across Japan, South Korea and China.
Directs, culls and hand-fixes machine-generated footage so it matches a production's style and holds continuity across shots. Emerging in advertising and previsualization since about 2024, and closer to art direction than to animation.
The honest reading is that this profession loses its bottom rung rather than its top. In-betweening, clean-up and colouring were both the automatable work and the apprenticeship, and removing them raises the entry bar without reducing the number of years it takes to become good. A studio that hires only mid-level animators for a decade will discover it has no senior ones.
Against that, the sheer volume of animation being commissioned keeps rising. Streaming, games, advertising, education, interfaces and virtual performers all consume animated motion, and demand for people who can decide what a character means is not obviously shrinking even where demand for people who can execute a tween has collapsed.
The likeliest 2040 animator supervises more output than one person could ever draw, works in tools that generate first passes, and is paid for exactly the thing that has defined the craft since 1908: knowing where to put the next position, and why.
The witness who finds out what is happening and tells the public — from Rome's Acta Diurna to the Panama Papers, still asking questions power would rather skip.
AI-resistant 58 📷The person paid to decide what a moment looks like — from Niépce's eight-hour exposure in 1826 to two trillion pictures a year, most taken by everyone else.
AI-resistant 55 🏅The body as a career: from Olympia's crowned victors to today's global stars, a short, brutal, winner-take-most profession watched by billions.
AI-resistant 88 🎭From Thespis stepping out of the Athenian chorus to the self-tape era: the professional pretender whose work is to make invented people feel true, take after take.
AI-resistant 70 🕹️From a Stanford lab's 1972 Spacewar tournament to sold-out stadium finals — the youngest, shortest and most-watched playing career in modern sport.
AI-resistant 74