The temperature inside the venue hovered near 40 °C, while an unprecedented heatwave scorched France outside. Yet animators, producers, and the financiers who bankroll and market their films filled the space to discuss the force remolding their field more swiftly than any prior generation: artificial intelligence.
Each June, the Alpine lakeside town of Annecy draws the eyes of the global animation community. This year, however, the sweltering heat was not the sole topic of conversation; artificial intelligence permeated every discussion, even as it remained largely unspoken.
On stage: the optimists’ case
The session, titled “Animation: More Human Than Ever,” was chaired by Mark Flanagan, a seasoned computer‑graphics instructor and founder of the VFX Jam training platform. He was joined on stage by Henry Daubrez, a Google Labs resident filmmaker; Jade Hautin, a producer at Paris‑based Frogbox; American technologist‑filmmaker Benjamin Michel; and producer Leo Neumann. Their discussion centered on a question that had drawn the crowd: how can animation retain its humanity when the tools that produce it are increasingly automated?
Daubrez championed accessibility, asserting that AI could finally equip creators in regions that have never possessed studios or professional software with powerful production tools.
He also cautioned about pitfalls. When applied without discernment, AI tends to default to the “average or the mean,” he said. The key, in his view, is to inject a distinct perspective into the machine rather than rely on it to generate one. His preferred approach — “hybrid production” — delegates rendering to AI while humans retain authority over motion and design.
Michel addressed economics, foreseeing a shift toward smaller studios — perhaps $5 million operations — producing films that previously required $50 million budgets, thereby forcing major houses to pare down what he termed “padding.” He repeatedly emphasized a insight that resonated throughout the room: “what’s left is you,” your taste, your eye. The dialogue repeatedly returned to questions of authorship, with one panelist noting that control is itself an act of creation.
Flanagan openly highlighted the paradox: seasoned filmmakers are attracted to AI because it could finally fund their passion projects that no one else would back, while younger artists worry about ever breaking into the industry.
Hautin, who has overseen two years of real‑world testing of these tools within her collective, voiced the ambivalence felt by many: “Part of you wants it to work,” she confessed, “and part of you doesn’t.”
Neumann was more direct about the efficiencies touted by others; for a small team, he argued, they would have moved faster without AI.
The panelists presented their most optimistic and most skeptical scenarios, ultimately converging on a point of consensus: no one can predict where this trajectory will lead in three years.
Off stage: the taboo
Beyond the tent, the discussion fell silent. AI has become animation’s open secret — embedded in virtually every stage of production — but acknowledging its use has grown nervy. Everyone aspires to be the first to achieve something groundbreaking with it, yet almost no one dares to openly admit they are employing it, fearing the repercussions.
The industry had just witnessed a similar episode. Weeks before the festival, Amazon MGM Studios and Amazon Web Services announced a fund to finance AI‑generated series, green‑lighting three projects for Prime Video. One of them, Punky Duck, was conceived by Jorge R. Gutiérrez, the Mexican director of The Book of Life and Maya and the Three. The response was harsh, extending beyond the use of AI itself.
Gutiérrez had long been one of animation’s most vocal advocates for junior artists; as recently as 2024 he warned that reliance on AI would remove the ladder that enables emerging talent to ascend, potentially stifling an entire generation of creators from producing hits.
Within two days of the backlash — which he said included threats to his family — Gutiérrez withdrew from Amazon’s AI initiative. “Actions speak louder than words,” he wrote, apologizing to anyone he may have disappointed.
Efficiency that wasn’t
Few have scrutinized such promises as directly as Leo Neumann, who leads a 30‑person German studio. In his film The Amazing Kitsuverse, he experimented with AI for tasks such as lip‑syncing and licensed voice work, only to emerge wishing he had avoided it. For a small team intent on retaining creative control, the overhead of evaluating and integrating these tools often consumed more time than it saved.
His concise ethical rule is simple: never surrender creative control, and never violate copyright. He goes further, likening the act of entering a prompt to hiring a stranger online; whatever the output, it remains, in his view, not truly yours.
He discovered the cost of transparency the hard way. When his studio listed every tool in the credits, an Annecy screening halted the moment AI was mentioned, whereas studios that remained silent continued undisturbed until their use was uncovered. He likens this to music that loses its value the instant listeners realize a machine, not a musician, is performing.
The missing first rung
For aspiring professionals, the concern is straightforward. Mexican animator Quique Gasca, a recent animation‑school graduate, worries that AI will first target the in‑between frames — the menial tasks that have traditionally taught juniors their craft and allowed senior artists to mentor them.
Eliminating that foundation removes the ladder entirely. Worse still, he argues, a model that has ingested everything already “has all the voices,” making the one thing newcomers crave most — a distinct voice — exceptionally difficult to discover.
His — and many of his peers’ — response is to double down on what machines cannot replicate: reviving stop‑motion techniques, embracing the tactile qualities of real materials, and celebrating the imperfections that arise from human hands.
What unsettles him is that, however appealing the handcrafted route may be, it remains a niche, while inexpensive AI‑generated “fast food” threatens to become the primary fare for the next generation.
The conversation comes apart
Jade Hautin, who also sat on the panel, observes from a different perspective. At Frogbox, where she produces, her studio avoids generative AI, and she notes that many French studios do the same. She additionally serves as an ambassador for Creative Machines?, a French‑language collective that employs a question mark in its name to signal its purpose: to interrogate the technology rather than market it.
What started in late 2023 as a handful of individuals exchanging links has grown into a chorus of over 1,100 members. They conduct sprints where participants test tools on genuine production work, dissect marketing hype, and host discussions with sociologists, lawyers, and economists.
What fascinates her most is the stark polarization surrounding AI discourse. Two camps have solidified: daily users who argue the industry must move forward with AI, and those who view any discussion of it as betrayal. Her collective attempts to occupy a middle ground, drawing criticism from both sides — being seen as too cautious by one, and a shill by the other. Meanwhile, the tools continue to advance; in April 2024 they could not even make an AI character blink, but by April 2026, she says, the results were astonishing.
She contends that this fear drives conversations underground. During the festival, her collective organized an international think‑tank of 40‑60 participants, and several attendees later told her it was the first venue where they felt safe discussing AI openly.
She argues that the true problem is not the assistive tools that have long resided in technical pipelines, but generative AI trained on scraped data, which she describes as having a monstrous environmental footprint. “Especially with this heatwave,” she adds.
What are we even talking about?
If the week clarified anything, it was that the industry cannot debate a term it has yet to define. Too much gets lumped under the label “AI,” Hautin observes.
Generative tools embedded deep within specialist pipelines differ fundamentally from the consumer‑grade applications anyone can summon with a browser click, and the former have been part of animation for years.
The real contention revolves around the latter category; models trained on works whose authors never consented to have their material used. She suggests that naming the technology precisely would enable the industry to argue honestly.
Outside, the heat continued its devastation. In a corner of the festival, a group of Spanish and Italian creators discussed precisely this issue: the environmental toll of the technology everyone else is debating. They love their work, they said, and it is their life.
But if the only path to a sustainable future for animation requires traversing an environmental catastrophe, then it is not worth pursuing. On that point, at least, they found common agreement.
