When Anyone Can Make a Movie
As artificial intelligence lowers the barriers to creating moving images, cinema faces a question more fundamental than how films will be made: what will make them worth watching?
The lights go down.
For the next two hours, we accept an arrangement that has existed for more than a century. Somewhere, actors once stood before a camera. Someone built the room behind them. A cinematographer waited for the light. A costume was fitted. Rain may have delayed the shoot. A hundred people may have stood just outside the frame.
We see none of this.
But perhaps, without realising it, we have always felt its presence.
Now imagine the same screen.
The actor was never born. The room was never built. The landscape has no coordinates.
There was no camera.
No one stood there.
And yet the film begins.
This is the threshold cinema is approaching as generative artificial intelligence moves from experimental clips into professional production. AI is entering workflows from pre-visualisation to post-production, while filmmakers, studios and performers negotiate what should remain distinctly human.
The questions are often economic.
How many jobs will disappear?
How much cheaper will films become?
How quickly can a scene be generated?
These questions matter. But they may not be the most interesting ones.
Because if the technology continues to improve, cinema may eventually confront a stranger possibility.
Making convincing moving images may no longer be particularly difficult.
And when anyone can make a movie, we may have to ask what a movie is for.
The physical infrastructure of traditional filmmaking: crew, lights, cables and equipment gathered just outside the frame. Pexels / Engin Akyurt
The Impossible Production
For most of cinema's history, imagination has had to negotiate with matter.
A filmmaker may imagine a city that does not exist, but someone must build it, paint it, photograph it or construct it digitally.
A director may imagine a storm, but the production must create rain.
A story may take place on Mars, but the actors still need somewhere to stand.
Cinema has always been a negotiation between what the mind can imagine and what the world can physically provide.
This limitation shaped the medium.
The camera needed light. Film stock had a cost. Sets occupied space. Actors aged. Weather changed. Budgets ended.
The image emerged from these constraints.
Generative AI begins to loosen that relationship.
The economic shift is already becoming visible. In July 2026, Netflix said that roughly 300 titles had used generative AI somewhere in production, primarily in post-production. For The American Experiment, the company said AI-assisted work allowed 17 minutes of complex imagery to be produced in roughly half the time and at half the cost of traditional methods, including imagery that might otherwise have been excluded for budgetary reasons.
This changes more than the price of an image.
It changes which images become possible in the first place.
A filmmaker may increasingly be able to describe rather than construct.
A street can be generated. A face can be invented. A performance can be synthesised. A location can exist for several seconds and disappear forever.
A 2026 study by Pierluigi Masai, Lorenzo Carta and Mateusz Miroslaw Lis argues that generative AI should not be understood simply as an assistant to existing filmmaking. The authors describe technologies capable of reconfiguring professional roles, production timelines and even film aesthetics. The distinction matters. A new camera changes how an image is captured. Generative AI may change which parts of the image need to be captured at all.
For independent filmmakers, this could be extraordinary.
Stories that once required enormous budgets may become possible for people who previously had little access to the machinery of cinema.
And the technical frontier is not confined to Hollywood. Researchers in China have developed systems such as FilMaster, an experimental end-to-end AI filmmaking framework trained to incorporate cinematic principles including camera language and editing rhythm. Its developers used a reference corpus of 440,000 film clips in an attempt to move generated video closer to the structural language of professional filmmaking.
The implications extend beyond any single system.
A filmmaker with a laptop may eventually command resources that once required a studio.
The distance between imagination and image becomes shorter.
This is the democratic promise of generative cinema.
But every removal of scarcity creates another form of scarcity somewhere else.
When Images Become Abundant
Human beings have always gathered around stories.
Long before cinema, stories were spoken around fires, painted onto walls, performed in theatres and carried through generations by memory.
The form changed. The gathering remained.
Cinema industrialised that ancient ritual.
Hundreds of strangers could sit silently in darkness and experience the same dream.
The cinema as collective ritual: hundreds of strangers agreeing to look at the same thing, at the same time, for a while. Pexels / Tima Miroshnichenko
The anthropological significance of cinema may therefore lie not simply in moving images, but in collective attention.
We agreed to look at the same thing. At the same time. For a while.
AI does not necessarily threaten this ritual.
Abundance does.
If generating a film eventually becomes as accessible as taking a photograph, the world may not suffer from too few movies.
It may suffer from too many.
Thousands become millions. Stories appear continuously. Every imagined world can be rendered. Every alternative ending generated. Every viewer could theoretically receive a film designed specifically for them.
The scarce resource would no longer be production.
It would be attention.
In that world, the most difficult part of filmmaking may no longer be making the film.
It may be persuading another human being to spend two hours watching it.
When images become infinite, attention becomes the final form of scarcity.
The Arrival of the Camera
Cinema has faced technological disruption before.
At the end of the nineteenth century, the motion-picture camera transformed still photography into something uncanny: time itself seemed capable of being captured.
L'Arrivée d'un train en gare de La Ciotat (1895). The Lumière brothers showed that ordinary life, simply recorded, could astonish. Wikimedia Commons / Public Domain
The earliest films were extraordinarily simple by contemporary standards.
Workers leaving a factory. A train arriving at a station. People walking through a street.
Nothing needed to explode. No fictional universe needed to be constructed.
The astonishment came from seeing ordinary life move after the moment had passed.
Cinema was new because the machine had learned to preserve time.
Over the following century, each technological shift appeared capable of changing the medium completely.
Sound arrived. Then colour. Television entered the home. Video made recording cheaper. Digital cameras replaced film in much of the industry. Computer-generated imagery allowed filmmakers to construct worlds that cameras could never photograph. Streaming removed the cinema from the centre of film distribution.
Yet cinema continued.
The technologies changed what could be made. They did not remove our desire to watch.
AI may follow this history.
But there is one important difference.
Previous technologies changed the tools used to record or construct performances.
Generative AI can potentially question whether the performance needed to happen at all.
The Actor Who Was Never There
A face appears on screen. It smiles. It hesitates. Its eyes fill with tears.
We understand the emotion immediately.
But suppose the person never existed.
Does that knowledge change what we feel?
This question is no longer entirely theoretical.
AI-generated digital humans created by NVIDIA. The faces were never photographed. They were computed. NVIDIA Technical Blog
Under the tentative 2026 SAG-AFTRA TV/Theatrical Agreement, producers must first bargain with the union over the use of synthetics. The framework establishes a principle strongly favouring human performance and states that producers do not intend to use a wholly synthetic performer in a human role unless it brings "significant additional value" beyond what could be achieved by a human performer or digital replica.
The effect is not a simple prohibition on synthetic performance.
It is something more revealing.
An industry is beginning to negotiate the conditions under which replacing human performance is considered valuable enough to justify.
In contractual language, cinema is already confronting a question the audience may eventually face for itself:
What does the presence of a real human being add?
SAG-AFTRA members on strike in Hollywood, 2023. The question of what AI can and cannot replace was already being debated on the street before it reached the contract. Editorial press photo
These debates begin with labour.
Who owns a face? Who controls a voice? Who is compensated when a likeness continues performing after the performer leaves the room?
Those questions are essential.
But another belongs to the audience.
What exactly are we watching when we watch a performance?
An actor does more than produce a convincing arrangement of facial expressions.
They bring a body. A history. Fatigue. Instinct. Embarrassment. Memory. The possibility of failure.
A performance contains the knowledge that another human being attempted something.
Perhaps that knowledge is invisible. Perhaps it is also part of what moves us.
The Aura of Something That Happened
In 1935, Walter Benjamin considered what technological reproduction might do to art.
Photography and film allowed images to travel beyond the singular location of an original work. In Benjamin's account, mechanical reproduction weakened the unique presence, or aura, associated with an artwork's existence in a particular time and place.
Yet reproduction also opened the work outward. An image no longer needed to remain where it was made. More people could encounter it. Art became more accessible even as its relationship with singularity changed.
AI extends this paradox one step further.
The question is no longer only what happens when an artwork can be reproduced infinitely.
It is what happens when new images themselves can be produced infinitely.
Access expands. Scarcity disappears. And value must migrate somewhere else.
Perhaps, in cinema, it migrates towards attention.
Cinema has always complicated the idea of an original. There is no single physical film experience in the way there is a single original painting.
The movie is reproduced every time it is projected.
And yet traditional cinema contains another kind of origin.
Something happened. A person stood somewhere. Light touched a face. A camera recorded it.
Even the most artificial blockbuster contains fragments of physical reality.
A hand. A costume. A breath. A performance.
Generative cinema introduces the possibility of an image without that event.
The photograph once seemed to say: This happened.
The generated image says: This could have happened.
The difference is small in language. On screen, it may become profound.
The Return of the Real
Perhaps this will eventually change how we value traditional filmmaking.
Photography did not eliminate painting. Recorded music did not eliminate concerts. Digital books did not eliminate paper. Industrial furniture did not eliminate craftsmanship.
Instead, abundance changed the meaning of the older form.
What was once ordinary became intentional.
Film photography returned partly because digital photography became effortless. When something becomes abundant, its opposite acquires meaning. Pexels
Film photography returned partly because digital photography became effortless.
Vinyl records survived partly because music became invisible.
Mechanical watches persisted after phones made them unnecessary.
A mechanical watch inspired by the Technics SL-1200 turntable. Analogue objects did not disappear when digital alternatives arrived. They acquired a different kind of value. Designboom / AndoAndoAndo
Perhaps cinema will experience something similar.
A future film might quietly announce:
Shot on location. Performed by human actors. Practical sets. Natural light. No synthetic performers.
Not because these methods necessarily create better art.
But because provenance may itself acquire meaning.
The physical production could become part of the story surrounding the film.
We may want to know that someone travelled there. That the actor was cold. That the house existed. That the rain was real.
Reality, once assumed, could become a creative choice. Perhaps even a luxury.
A New Kind of Filmmaker
None of this means AI cinema will necessarily be inferior cinema.
That conclusion would be too easy.
Every major technological change in art has been accused of diminishing what came before.
The camera was once considered mechanical. Film was once considered inferior to theatre. Digital cinematography was dismissed by some for lacking the texture of celluloid.
New tools eventually develop their own language.
AI may do the same.
The most interesting AI filmmakers may not be those who use machines to imitate Hollywood more cheaply.
They may be those who discover images that traditional cinema could never have produced.
Dreams without physics. Memories that change while we watch them. Characters whose worlds respond to individual viewers. Films without fixed endings.
Cinema may become less like recording a world and more like growing one.
The mistake would be to judge this future only by whether it can perfectly imitate the past.
A synthetic film does not need to pretend that a camera was there.
Its artistic possibility may begin precisely when it stops trying to.
The production set: what if this visible labour — the crew, the equipment, the physical presence — becomes part of the value of the film itself? Pexels
The Last Human Film
Perhaps the future will not divide neatly between AI films and human films.
Most cinema may exist somewhere between them.
Human-written stories with generated environments. Real actors inside synthetic landscapes. AI-assisted editing. Digital replicas used with consent. Practical photography altered by generative tools.
The boundaries will blur.
And perhaps the question "Was this made by AI?" will eventually feel as incomplete as asking whether a contemporary film was made with a computer.
The more meaningful question will remain older.
Did someone have something worth saying?
Technology can make images cheaper. It can make production faster. It can place impossible worlds within reach of people who could never afford to build them.
That possibility should not be dismissed.
But technology cannot guarantee that anyone will care.
The cinema goes dark.
An image appears.
Perhaps there was a camera. Perhaps there was an actor. Perhaps there was only a machine producing pixels from language.
For a moment, none of this matters.
Then something happens on screen.
And we decide whether to keep watching.
That decision may become the true test of cinema in the age of artificial intelligence.
Not whether a machine can make a movie.
But whether, surrounded by an infinity of images, a movie can still make us feel that these particular images were worth our time.