Werner Herzog Says AI Filmmakers Should Fail—Then Explains Why

The reporting does not say that Herzog pursued legal action or tried to stop the film’s release. It says he declined participation and condemned the result. Keep the facts where they are. The insult is already doing plenty of work.

Photos: Colleen Sturtevant · CC BY-SA 4.0 · source; Fronteiras do Pensamento · CC BY-SA 2.0 · source

Herzog’s target expands from one project to an industry habit

In the interview, Herzog widened his attack beyond Winiewicz. He argued that AI can step in when filmmaking has become standardized and storytelling has disappeared. His description of the current failure is repetition: mimic what happened before, repeat it again, then repeat it until the audience is numb.

He even welcomed AI replacing “all the idiots” devoted to that cycle. That sounds like a provocation because it is one, but the point has a clean edge. If the work contains no discovery, why defend the human labor behind it? Herzog’s answer is that the machine may as well perform the imitation. At least then the production would stop pretending that repetition equals invention.

This is the reversal at the heart of his position. Herzog is not claiming that humans possess a magical copyright on every image or sentence. He is saying that AI exposes a human weakness already present: a film assembled from familiar gestures can be automated because its makers have already automated themselves.

Photos: Fronteiras do Pensamento · CC BY-SA 2.0 · source; Greg2600 · CC BY-SA 2.0 · source

Why human actors remain the part AI cannot replace

The material does not report a separate Herzog statement about actors, performance capture, or labor contracts. It does not establish a technical limit that makes synthetic performers impossible. So the argument here has to stay with what Herzog actually said.

His complaint points to a human element that imitation cannot generate on command: a person making an unrepeatable choice. An actor brings timing, resistance, accident, embarrassment, appetite—the small disruptions that keep a scene from becoming a quotation of another scene. A system trained on past work can reproduce the surface of those choices. The reporting gives no evidence that it can originate the risk.

That is why About a Hero lands as more than a failed experiment. Its premise turns a filmmaker into a database and calls the database a voice. Herzog’s answer is to hand the project its own logic: go ahead, imitate, repeat, and fail in public.

The sharpest image left by the reporting is not a robot taking an actor’s job. It is a machine stepping into an empty chair while the humans nearby keep applauding the furniture.

Every factual claim above is drawn from these. The reading of them is ours.