In late 2023, the Animation Guild commissioned an independent economic analysis of what generative AI would do to jobs in film, television, and animation. The result, produced by CVL Economics, projected that 21.4% of those jobs, roughly 118,500 positions, would be consolidated, displaced, or eliminated by 2026. That is not a distant projection anymore. It is close to the year this article is being published in.
The displacement risk for animators is classified as Moderate, with a horizon toward 2029, one of the nearer-term horizons on any Moderate-risk profession AI Doomsday tracks. That timeline is not speculative. Pixar cut 175 jobs in 2024, its largest reduction ever, and followed with further rounds in 2025 and 2026. DreamWorks Animation cut roughly 70 positions in 2024. Rooster Teeth and its animated content division were shut down entirely that same year. These are not projections of what AI might do. They are line items in earnings reports of what has already happened.
This is a different story from the one AI Doomsday told in its coverage of actors, which centers on performance, likeness, and deepfakes. Animation's exposure runs through production itself: the frame-by-frame, layer-by-layer labor of turning a keyframe into finished motion, and a growing share of that labor is exactly what generative tools now do passably well.
Key Points
- Animators are rated Moderate risk with a displacement horizon toward 2029, and unlike most professions on this list, significant job losses have already occurred rather than remaining a future projection.
- A 2023 study commissioned by the Animation Guild projected that 21.4% of film, TV, and animation jobs, roughly 118,500 positions, would be consolidated, displaced, or eliminated by 2026.
- Pixar, DreamWorks Animation, Rooster Teeth, and Paramount's Nickelodeon and Cartoon Network divisions have all cut staff since 2024, in a period that overlaps directly with generative AI's entry into production pipelines.
- AI tools including Cascadeur, Adobe's Roto Brush and Object Matte, and Autodesk Flow Studio are already automating in-betweening, rotoscoping, and motion-capture cleanup in real production, not just demos.
- The Animation Guild's 2024 contract added AI disclosure and consultation requirements but no opt-out for animators, no ban on training models on their work, and no minimum staffing guarantees, a gap members publicly criticized before ratifying it 76.1% in favor.
What an Animator Actually Does
The title covers a production pipeline with dozens of specialized roles: character riggers who build the digital skeletons a character moves on, key animators who define the pose at each major beat of an action, in-betweeners who fill the motion between those poses, background painters, compositors, and motion-capture technicians who clean up raw performance data into usable character animation, all reporting up to an art director who sets the visual vision the whole pipeline serves. Historically, this pipeline has been structured like an assembly line precisely because each stage, however skilled, is repeatable enough to divide across a large crew.
That structure is also what makes the pipeline unusually exposed. A job built from clearly separable, high-volume stages is easier to automate piece by piece than one where a single person owns an entire unpredictable task from end to end.
What AI Is Already Doing
The in-betweening stage, filling the motion between two key poses, is furthest along. Japanese animation studios, facing both AI capability and an acute labor shortage, at Toei Animation and elsewhere, have adopted AI in-betweening and background generation tools that studios report cutting background art production from roughly a week to a matter of minutes. That adoption is not purely enthusiasm for the technology. Japan produced 310 anime titles in 2024 with fewer than 6,000 trained animators available, many earning under $1,400 a month while working roughly 219 hours monthly, well above the national average. AI is entering an industry that was already running on unsustainable labor conditions.
In Western production pipelines, Cascadeur uses AI-assisted physics simulation to auto-generate plausible in-between poses and interpolation from a small number of keyframes, in active use across game studios and independent film. Adobe's Roto Brush and newer Object Matte tools, now standard in After Effects, automate rotoscoping and masking that used to require frame-by-frame manual tracing. Autodesk's Flow Studio, built from its 2024 acquisition of Wonder Dynamics, uses AI to insert fully rigged CG characters directly into live-action footage with automated motion capture, a task that previously required a dedicated mocap cleanup team.
THE FORECAST
The Animation Guild's commissioned study did not just estimate a national number. It broke the disruption down by region: roughly 39,500 affected positions in California and 15,100 in New York, the two production centers where the bulk of unionized American animation work is based. The study was published in late 2023. Its three-year window closes around the release of this article.
Generative video tools, Runway, Kling, and the now-discontinued standalone version of OpenAI's Sora, have improved sharply on raw output quality, but remain a step behind what serialized production actually requires. Technical assessments of these systems consistently flag temporal consistency failures, a character's face, proportions, or costume drifting subtly across a single generated clip, as the core unsolved problem preventing their use for multi-shot, narratively continuous production rather than short, standalone clips.
Hollow Protections
The Animation Guild, IATSE Local 839, ratified a new contract in September 2024 that added specific AI provisions for the first time: studios must notify the union when generative AI tools are used and consult on alternatives. What the contract does not include is notable by its absence. There is no opt-out for individual animators whose work informs an AI system, no restriction on training models on union members' existing output, and no minimum staffing requirement tied to AI adoption. Three members of the guild's own negotiating committee voted against the deal, and working animators publicly described the AI language as protection in name only. It still passed, 76.1% in favor, a reflection of how much leverage a union has when the disruption it is negotiating against is already underway rather than hypothetical.
The same anxiety shows up industry-wide. GDC's 2026 State of the Game Industry survey, covering visual and technical artists alongside programmers and designers, found 52% of respondents describing AI's impact on the industry as negative, up from 30% the year before and 18% the year before that, with 64% unfavorable specifically among workers in visual and technical art roles. That trend runs parallel to what AI Doomsday covered in the broader game industry, where the same generative tools reshaping animation are compressing technical art headcount on the game side of the same production pipeline.
How to Use AI as an Animator Now
The animators best positioned right now are treating these tools as a way to move up the pipeline, not a threat to avoid.
For in-betweening and cleanup: Cascadeur and similar tools handle interpolation competently enough that fighting their adoption is a losing position. The higher-value skill becomes directing and correcting AI output, defining strong keyframes and physical intent that automated interpolation actually executes well, rather than doing the interpolation by hand.
For rotoscoping and compositing: automated masking tools remove the most tedious part of the job without removing the compositor's judgment about what the shot actually needs. Treat the automated output as a fast first pass that still requires a trained eye before delivery.
For character integration: Autodesk Flow Studio and comparable tools compress what used to be a specialized mocap-cleanup role into a faster pipeline stage. The animators who stay relevant are the ones who understand performance and motion well enough to art-direct the automated output, not just accept it.
What I Think
The 2029 horizon feels tight to me in the way it should. This is one of the few professions on this list where the disruption is not a forecast competing with uncertainty about whether the technology will actually arrive. The technology already arrived, the layoffs already started, and the union's own commissioned study is now inside its own prediction window rather than ahead of it.
What troubles me most is not the automation of in-betweening or rotoscoping. Those were always the most mechanical parts of the pipeline, and reasonable people can disagree about whether losing them is a tragedy or an overdue efficiency gain. What troubles me is the pattern visible in Japan: an industry already extracting unsustainable hours from animators at poverty wages, adopting AI not primarily to lower costs further but to survive a labor shortage its own working conditions created. AI is not arriving in animation as a disruptor of a healthy industry. It is arriving as a patch for an industry that had already broken its own workforce, and a patch is not the same thing as a fix.
"The union got a seat at the table and a right to be told. It didn't get a right to say no. In an industry where the disruption study and the layoffs are running on the same three-year clock, that gap is the whole story."