What is the purpose of a screenplay?

A screenplay serves two functions: to convince and to organize. AI challenges this transitional format by enabling more flexible and interactive creation, without eliminating traditional methods. Paths for experimentation, to renew cinema, so that it remains alive.

Two roles

The screenplay is a preliminary object for the creation of a film, which has two main functions:

1. To convince.
2. To organize.

The screenplay serves to convince

The screenplay must convince its readers to engage in the adventure of this film. To engage as an actor, as a producer, as a financier, as a distributor, as a technician, etc.

This is why the screenplay is formatted. One of the reasons for formatting the screenplay into scenes, one minute per page, with a reading time that takes as long as the film, is that the screenplay is a very synthetic object; it’s not at all an object that tells everything that will happen in the film, otherwise it would be much longer. So this object, which allows one to imagine the film by reading a simple written document that costs much less to produce than an entire film, allows for convincing.

The screenplay serves to organize

The division of the screenplay into scenes, the precision of the characters present in each scene, which corresponds to a specific filming location, gives the screenplay this function of a technical tool, which will be the basis from which we will organize the shooting of the film, that is to say, contacting people, building sets, finding props, preparing the necessary technique and teams to be able to implement things, respect the budget, etc.

On the transitional object

If the screenplay has this transitional place, it’s because it costs less to produce than the film. If we could directly make films, the screenplay would no longer have the role it has today. It has this role because it’s expensive, it requires mobilizing many people, and therefore we need to find the means and organize well to be able to gather the human beings necessary for the making of the film. Whereas after all, writing on paper can be the work of a single person or a team of people, but it doesn’t mobilize all the money, technique, and all the collective work time necessary for the making of the film.

It is also very interesting to question the social statuses of the different participants in the making of a film. The screenwriter is not an intermittent performer, he has an extremely precarious social status. He is free, he is not subordinated to anyone, he ultimately creates the source. Whereas for the making of the film, all the people are salaried, receive unemployment benefits if they are laid off, etc. It’s a technical team. The director himself is a technician, legally speaking.

The screenplay is part of a system for organizing matter, space, time, and resources.

Artificial intelligence

But if, today, in 2025, they are not yet extremely developed, there are generative artificial intelligences that are capable, from “prompts,” that is to say micro-scenarios, of creating images and sounds, thus films or pieces of films or elements that serve to make films without the need for financing. These are almost infinitely capable technicians. These tools cost a little bit of money, but much less than shooting films with set construction, salaries of people, etc. So this type of practice for making films completely changes the role of the screenplay. Because if we no longer need to convince people because the actors, sets, and entire film are generated by artificial intelligence, nor gather very large sums to organize a shoot, finally, we are no longer obliged, if we use artificial intelligence, to write our entire screenplay to prepare for the shooting of the film.

We could, for example, progressively write the film, modify everything all the time, because finally, it’s only calculation time, but due to the power of machines today, we can quite easily redo a sequence several times. There is no longer even a need for the final step of the editor because the machine itself does the editing. We could, for example, on the final film, an hour and a half long, ask the machine that the whole film be a flashback and that the main character, instead of being the man in the couple, be the woman in the couple, and this would modify the whole story and the love relationships, and even the stakes, and thus bring to the film a political dimension on the question of gender that the first version did not have.

Today, early 2025, we cannot yet ask for this, we would not have an immediate result on a film of an hour and a half. But that will come quite quickly. The machine can therefore completely rewrite the screenplay and completely redo the film with these changes. It will be accessible.

In the same way that today, and this already exists, there are interactive podcasts, that is to say that instead of listening to a podcast, one can interact with the podcast. It’s an artificial intelligence that has all the data of the interlocutor, and one can ask questions. This interlocutor, the artificial intelligence, will draw from its database and will give answers that are relevant to who the interlocutor is. This creates another form of podcast, more interactive. One can choose the duration of one’s podcast according to the time one has. It’s no longer “Ah, I didn’t listen to the end, so I missed something.” I can go ask and immediately get what I’m looking for, orally.

At heart, even my conception of what a film is is quite old-fashioned. Indeed, I envision the film as a fixed object, which would not be interactive at all. The film as a fixed object has its necessities, because it costs so much to make that making many versions of the film would be complicated and very expensive.

Let’s remember that at the very beginning of cinema, it was much more complicated to make copies of the same film. The film “The Train Attack,” The Great Train Robbery, is a film of which there are many, many, many, many, many versions because, in fact, we shot the film, and it was the unique copy of the filmed movie that was shown. So, if several copies were needed, several shootings of the film were done!

Cinema and video games

One could object to me that interactive audiovisual objects already exist, they are video games. Will the future of cinema be to be a vast video game, if we compare to the podcast I just talked about? I don’t think so. I think we shouldn’t be afraid of technological evolutions. I think we’re adding brushes to the palette with the advancement of technologies. We see it clearly with regard to the book.

With the arrival of digital, there were many predictions about the future disappearance of the book. And in fact, it didn’t happen. There is a complementarity. Books have not disappeared, and we can, if we prefer, read books digitally. But there are no fewer books than before. And simply, there are more readers and there are more ways to read. So it’s ultimately very rich.

Thus, I believe there is no fear to have with artificial intelligences. There will always be screenplays and real film shootings. But there will also be more and more films made differently.

Films without a screenplay

So, there will be films that will be made without a screenplay, in cooperation with artificial intelligences, which will resemble films made with a screenplay, but which will have been co-constructed with artificial intelligences. I think these new working methods that will be invented will be very enriching for cinema, because perhaps the screenplay is also a brake on creativity in a certain way, because it’s a way of doing things with a certain number of steps, and perhaps by employing another method, not of writing, but of making films, we will discover new narrative modalities, new ways of telling stories thanks to these new techniques, which will undoubtedly then have influence even in the writing of “classic” screenplays.

Working differently to do something different

From today, I believe it can be interesting to experiment with other ways of making films than the traditional way “screenplay writing + film making,” which obviously imposes small budget films, and a priori shorter films. It’s very rich for cinema to experiment with new ways of doing, to risk it and to risk doing things that are not within the norms.

Artificial intelligences have an unfortunate bias: of course we will be able to invent things with them, but generative artificial intelligences rely on the corpus of pre-existing films. And today, this will undoubtedly change, but today, they have very little imagination. It’s quite impressive, because it immediately creates images, but in reality, these images are referenced to the images that these artificial intelligences know. Like us, in a certain way. But we still have in creation something more organic, less mechanical than artificial intelligences. So I think that can bring us a lot. But I see that there is also intrinsically a formatting.

Just as there is a form of formatting in the way of making films, which is to write a screenplay and then shoot the film, because it’s a method. And finally, making films with artificial intelligences will be another method, not better, nor worse, simply another method, but undoubtedly relatively formatted as well, even if it can still open doors.

That’s why it seems to me that it is interesting and important to experiment with other ways of doing, with and without artificial intelligences, to renew cinema. What’s the use of renewing cinema? It serves to contribute to the world in an enriching way for human beings, to open the doors of the imaginary, to share them, to build common references, thus to contribute to the construction of the world, ultimately.