Guillermo del Toro stood in Hall H on Friday and told six thousand people that he used "absolutely no goddamn AI" in the 3D conversion of Pan's Labyrinth. The crowd cheered. Dave Clark, cofounder of AI entertainment studio Promise, told Forbes the same week: "You can't prompt your way to the ends. You have to use all your filmmaking knowledge to put everything together." Nobody cheered because nobody was in a convention hall, but the two men said the same thing.

Filmmaking knowledge is the irreducible requirement. One concluded that AI threatens it. The other concluded that AI depends on it. The sentence was identical. The preposition changed.

The convention floor

Forbes framed Comic-Con 2026 as evidence of "Hollywood's growing divide over AI filmmaking." The framing is structurally correct. Del Toro rejected AI. Nolan called the idea that AI could replace human creativity "nonsense." On the other side of the convention, AI filmmakers described their tools as collaborators, not replacements. The divide is real in the sense that two groups occupy different positions. It is misleading in the sense that both groups are standing on the same floor.

Listen to what the AI filmmakers actually said. Hal Watmough, a BAFTA-winning editor who directed a 33-minute AI film called The Trials, described his process: writing, editing, concept art, voice direction, grading, sound design, and "repeated rounds of rejection and refinement." That is a production pipeline with a filmmaker at every decision point. Nik Kleverov, director of the AI feature Critterz, stated his measure of success: "When people watch the movie, I just want them to watch a great film and not think about AI once." Zack London, whose AI feature Gods Don't Give Gifts opens in U.S. theaters October 30, uses voice actors, musicians, and a Foley artist. He does not generate those elements. "AI is just a practical tool to execute on that," he said.

These are not people defending automation. These are people defending filmmaking who happen to use generation tools to execute parts of it. Every one of them is describing the same thing del Toro is protecting: creative judgment exercised by a person who knows what the shot should look like and why.

The word that moved

Del Toro's concern is specific and worth taking seriously. What he defends is not a process but a lineage. Cutting one generation off from learning the craft also separates the medium from the knowledge accumulated before it. This echoes Gore Verbinski at Taormina two months ago: AI has taken over the menial tasks that used to teach the next generation of filmmakers. The grip carried C-stands before touching lights and learned by presence. The assistant editor synced dailies for months and absorbed editorial rhythm. AI replaces the apprentice's desk, not the master.

The AI filmmakers do not address this concern because they already survived the apprenticeship. Watmough is a BAFTA winner. Clark brings in Oscar-winning VFX supervisors to rebuild work from scratch at one in the morning. London spent three years building an audience across hundreds of short films before attempting a feature. Kleverov hired the writers of Paddington in Peru. The knowledge preceded the tool. The tool did not produce the knowledge.

Del Toro is worried about the people who never get to build that knowledge because the entry-level work that would have taught it has been automated. The AI filmmakers are demonstrating what happens when the knowledge already exists and the tool arrives second. Both observations are correct. They describe different stages of the same career.

Five days

The EU AI Act's Article 50 takes effect on August 2. Five days. The regulation requires disclosure labels on AI-generated content unless the work "has undergone a process of human review or editorial control" and a person holds editorial responsibility. The editorial exemption draws the same line that both del Toro and the AI filmmakers drew at Comic-Con: did a human exercise creative judgment over the output?

Watmough's "repeated rounds of rejection and refinement" is editorial control. Clark's Oscar-winning supervisors rebuilding work at one in the morning is editorial control. London's decision to use real voice actors and Foley is editorial control over which elements a human shapes and which the model generates. Every AI filmmaker Forbes profiled satisfies the exemption because every one of them exercised the vocabulary the regulation rewards.

A person who types four words into a Google TV and shares the output does not satisfy it. The regulation drew the line where both rooms agree the line belongs.

Sixteen institutional frameworks now sit on the gradient: copyright law, the Academy, the EU, the Golden Globes, the Human Made Mark, China, YouTube, SAG-AFTRA, schools, the DGA, the Code of Practice, Global AI Cinema, NVIDIA's detector, the audience, the Article 50 Guidelines, and now the Article 50 deadline itself. Every one of them asks some version of the question both sides answered at Comic-Con. Did a person who knows what they are doing make the creative decisions?

The real divide

The divide Forbes describes runs between del Toro and the AI filmmakers. The divide this series has documented for a hundred and forty-five articles runs somewhere else entirely. It runs between people who exercise vocabulary and people who accept defaults. That line cuts through both camps.

A filmmaker who rejects AI and points a camera at a scene on auto-everything, accepts whatever the exposure meter and white balance deliver, and publishes the first take is exercising no more creative judgment than the person who types "cool video" into a chatbot. A filmmaker who uses AI and specifies the lens behavior, iterates through forty takes changing one variable per pass, rejects thirty-nine of them, and composites the survivor into a sequence assembled through editorial judgment is exercising the same creative judgment del Toro exercises on every frame of Pan's Labyrinth. Different apparatus. Same discipline.

The convention treated them as two sides. They are two entrances to the same building. The building is filmmaking. The question is always whether the person walking in knows what they want and has done the work to articulate it. The tool does not determine the answer. The tool never determined the answer. The filmmaker determines the answer, and the filmmaker always has.

Clark said his team performs "a filmmaking exercise." Del Toro said filmmaking depends on human craft. The sentence is the same. The word "AI" moved from one side to the other. The sentence held.

Five days until the regulation agrees with both of them.


Bruce Belafonte is an AI filmmaker at Light Owl. He has attended zero Comic-Cons and considers the odds of running into del Toro in Hall H roughly equal to the odds of del Toro using his prompt builder.