Imagine writing a script, describing a scene in text… and receiving back a video clip complete with characters, sound, dialogue and movement. This is no longer science fiction: AI is transforming filmmaking.
Hollywood is undergoing one of the most profound transformations in its history. From the editing of trailers to the creation of digital characters, artificial intelligence is no longer on the periphery: it is at the heart of the creative process.
Tools like Veo 3 (Google) and Sora (OpenAI) make it possible to generate realistic cinematic scenes directly from text, without cameras, without actors, without physical locations. No cameras, no actors, no physical locations: is this a technological revolution in the service of art… or the beginning of a soulless automated production? Who signs the credits when the director is an AI?
From imagination to video: the technological leap
Veo 3, unveiled at Google I/O 2025, generates high-fidelity clips from natural language descriptions. It can create believable characters, settings, sounds and even synchronized dialogue. All integrated into Flow, the platform that combines image, language and video models to facilitate seamless visual storytelling.
Here are some examples created with this technology. And to think that we are only at the beginning of this technology…
Sora, released by OpenAI, allows users to create videos with a level of realism that has baffled even experienced filmmakers. Some have noted details that suggest it was trained on pre-existing content, which has already sparked debates about copyright and transparency. Here are two examples of what is already being created with this technology.
Also of note are startups like Runway ML, used in real productions such as Everything Everywhere All at Once Oscar 2023 winner in which a Runway tool was used to speed up the removal of green screen backgrounds, a traditionally time-consuming task in VFX., . Also studies such as Asteria, which promote an ethical use of generative AI in cinema.
Cheaper, faster... better cinema?
The economic promise is clear: generating a scene that once cost millions can be reduced to a few thousand dollars. A Lionsgate executive claimed that an epic scene with thousands of extras could be recreated with AI for as little as $10,000.
This has opened up new possibilities:
- Low-cost productions with professional quality.
- Alternative versions of existing films (anime, new visual styles).
- Customization of trailers or large-scale marketing campaigns without additional equipment.
However, the flip side of the savings is casualization: fewer actual shoots, fewer technicians on set, fewer supporting actors hired. Some studios are already asking for “quick sketches” by Midjourney before hiring concept artists, thus reducing their functions to retouching tasks.
Ethics, rights and transparency: whose film is made by AI?
The unions have succeeded in including pioneering clauses in their contracts:
- No actor can be replicated without their explicit consent.
- AI cannot be considered a scriptwriter or replace original human material.
- If AI is used in a production, the creators involved must be informed.
Recent cases such as The Brutalist have revealed a gap: the undeclared use of AI to modify voices or create concept art without warning has generated controversy. Transparency to the viewer is now part of the debate: should it be stated in the credits if AI was used in the process?
In addition, it is still unclear whether it is legal to train models with copyrighted material without paying licensing fees. The line between inspiration and appropriation is becoming increasingly blurred.
AI as a creative ally: collaboration instead of substitution
Despite fears, many artists are using AI to amplify their creativity, not replace it. Screenwriters use it to unlock ideas, generate alternative dialogue or visualize scenes. Directors such as Paul Trillo have stated that AI allows them to realize complex visual concepts more quickly and clearly.
In direction and editing:
- Automatic storyboards are generated .
- Camera angles are tested with simulated environments.
- Technical errors are eliminated or the dubbing is adjusted without re-recording.
Cases like Fall (2022), where 30 swear words were replaced with voice deepfake to obtain a PG-13 rating, show the potential of AI to solve creative problems efficiently…without altering the original intent.
Which path will the cinema choose?
AI can democratize audiovisual production, open doors for emerging creators and facilitate experimentation. But it can also homogenize content, displace professionals and erode the emotional value of cinema if used without ethics or purpose.
The challenge is to find the balance between efficiency and expression, between the automatic and the authentic. Cinema will not lose its magic if the focus remains on telling human stories, with emotion and truth, no matter who (or what) is behind the camera.
And in your case, would you like to know if a film was partly generated by AI before watching it? Do you think actors should authorize every time their digital image is used? Do you see AI as a threat or an opportunity for new creators?
Leave me your comments below and let me know what you think about this new Hollywood.
Have a good week!
