EDITION→
Global
•
USA
•
UK
•
AUNZ
•
CANADA
•
IRELAND
•
FRANCE
•
GERMANY
•
ASIA
•
EUROPE
•
LATAM
•
MEA
Edition : InternationalLanguage : English
HOME
NEWS
CREATIVE LIBRARY
IMMORTAL AWARDS
MEMBER COMPANIES
PEOPLE
JOBS
How LBB WorksPlans & PricingEditorial Toolkit
news
NEWS
CREATIVE LIBRARY
IMMORTAL AWARDS
MEMBER COMPANIES
PEOPLE
JOBS
Login or Register
People in association withThe Monthly Cut
AI Filmmaking: Balancing Creativity, Craft and the Human Element with Pablo Díaz
lbbonline.com
0
LIKE
ADD TO COLLECTION
28/09/2026
Share
The editor at SOMOS5 on the advantages and disadvantages of a production made predominantly with AI, and discovering the best tools to use, as part of LBB’s AI Spy series
Pablo Díaz has built his path in the audiovisual field by learning to navigate different languages without ever losing his own voice. He began in a highly self-taught way, founding his own production company to create the opportunities he couldn’t yet find in the market and to gain hands-on experience. A director, editor, and creator, his career spans advertising, television, digital content, and fiction, always with an approach that seeks more than just beautiful images, but pieces that resonate emotionally.
His profile combines the precision of an editor, the sensitivity of a storyteller, and a broad vision of the audiovisual project. He has worked on advertising campaigns, documentaries, music videos, television formats, and digital content, as well as coordinating teams. In his way of working, craft experience, intuition, and a creative sensitivity coexist – one that goes beyond simply solving problems, and instead aims to leave a lasting impression.
In parallel, he has developed a more authorial side in cinema, directing and writing short films recognised at national and international festivals, while also contributing as an editor on two feature films. This dual position – between industry and personal creation – defines his professional identity: someone who knows how to find the right tone for each piece and bring it to that point where technique and emotion meet.
Pablo sat down with LBB to discuss AI transformative potential in post production, the dangers of generative AI, as well as his thoughts on hybrid productions.
LBB> In which specific stage of your commercial filmmaking process (pre-production, on set, or post) is AI currently delivering its most transformative creative or logistical results, and what specific tool do you rely on most?
Pablo> At the moment, the most transformative impact of artificial intelligence in my own process is in pre-production, particularly in the creation of visual treatments, aesthetic development, and previsualisation.
AI dramatically reduces the distance between an abstract idea and its visual representation. I can develop characters, locations, atmospheres, framing, or lighting tests before hiring the crew or stepping onto a set. Does this replace creative decision-making? No, but it allows those decisions to be made with far more information and the vision to be communicated much more precisely.
My go-to tools at the moment are Magnific and Higgsfield. I use them as central platforms for testing different models and combining image generation, video generation, and visual enhancement workflows. The specific model changes depending on the needs of each project.
I also use AI in post-production for tasks such as shot clean up, set extensions, or generating specific elements. However, where I find its greatest transformative potential is in pre-production: it allows us to visualise, test, and discard ideas at a speed that was previously impossible.
LBB> How has AI fundamentally changed how you build a visual treatment or pitch a concept? How do you sell a vision to an agency or brand when the most ambitious visuals can be generated digitally before a single crew member is hired? Does it make it easier to sell in an approach or does it make it harder to stand out?
Pablo> AI has turned the visual treatment into something closer to a prototype of the film. At Somos5, we previously worked mainly with references taken from other films, photographs, or campaigns. Now we can generate images specifically for a project: characters, locations, compositions, and atmospheres that come much closer to what we actually want to shoot.
This makes it easier to sell an approach, because it reduces the amount of imagination required from the agency or client.
However, it also makes it harder to stand out. The ability to generate visually striking images is becoming increasingly democratised. As a result, differentiation can no longer rely solely on the spectacle of an image. It has to come from the idea, the point of view, the visual coherence, and the ability to turn those images into a narrative.
LBB> When it comes to AI in filmmaking, what are the pros and cons of fully-AI productions versus hybrid productions?
Pablo> First of all, I think we need to qualify the idea of a ‘100% AI’ production. Even in a film generated almost entirely through artificial intelligence, there is still human involvement: in the idea, the script, the instructions, the references, the selection of outputs, the edit, sound, colour, and budgetary decisions. Ultimately, it comes down to judgement. Without that series of decisions, there is no film.
The main advantage of a production made predominantly with AI is that it can reduce or eliminate much of the logistics involved in a shoot. From a computer, you can create locations, characters, and situations that, using traditional methods, would require large budgets, travel, or complex technical crews. It also allows you to test versions quickly and explore routes that would probably never be attempted in a conventional production.
Its main drawbacks are still control, consistency, and human performance. Models can produce artefacts, unintentionally alter characters or elements, and deliver unpredictable results. And although control is improving, it is still harder to specify a precise action, position, or physical relationship than it is to direct it on set.
You also lose part of the collective creative process. A shoot is alive. Every department interprets and enriches the film through its own expertise. That collaboration can generate results that a director working alone in front of a machine would be unlikely to anticipate.
That is why we currently favour hybrid productions. AI can streamline processes, extend environments, generate elements, and enrich a piece, while traditional filmmaking continues to provide performance, precision, and human experience.
LBB> In terms of the production process, what elements of the traditional production process do you see mirrored in the AI filmmaking workflow (ppms? Casting? locations?) and in what ways does the new way of working differ wildly? Is there a ‘standard’ AI production workflow or do you find you need to build anew for each project?
Pablo> The processes involved in a generative production are quite similar to those of a traditional production, even if the language and tools are different.
We still begin with a brief or a script. We then break down the shots, develop the characters, define the locations, wardrobe, art direction, props, and overall aesthetic. The difference is that, rather than physically finding or building those elements, we first have to create and lock them as digital assets.
In an AI production, developing a visual bible is particularly important: clear references for characters, locations, proportions, materials, palette, lighting, and camera language. The more coherent information we give the model, the greater our chances of maintaining continuity.
Our usual workflow at Somos5 starts with the script and shot breakdown. We then develop the look, characters, and locations. Once those elements are approved, we generate the still images for each shot. We then animate them to achieve the required camera movement and actions. Finally, the piece moves into editing, sound design, and colour for the final master.
LBB> AI is creating new roles (prompt engineer, AI artist). How is your production company restructuring or upskilling your teams to handle this “hybrid” filmmaking? What does the ‘crew’ on a AI production project look like?
Pablo> I don’t think this evolution is simply about bringing in someone who knows how to write prompts. A prompt is only a very small – albeit essential – part of the process. What really matters is understanding image-making, narrative, continuity, and the possibilities and limitations of each model.
On a small generative production, the crew might consist of a director or creative lead, an AI artist specialising in image-making and visual development, someone specialising in animation or video generation, a post-production team, and a producer. Depending on the project, specialists in sound, colour, visual effects, and legal supervision may also be involved.
In a hybrid production, these roles should not necessarily replace traditional departments, but integrate with them. A cinematographer can use AI to develop lighting and the visual approach to a shot; a production designer can use it to explore spaces and materials, etc.
LBB> The speed of new releases means that there’s a different set of ‘best’ tools every couple of weeks! How do you keep up with that and what are your go-to tools right now? To your mind, what have been the most significant recent developments?
Pablo> I try not to chase every new release. A new model appears practically every week, but that doesn’t mean they all bring a meaningful improvement to the work. I prefer to test them through specific exercises and compare factors such as control, consistency, movement, speed, cost, and commercial-use terms.
In this field, it is important to be self-taught because knowledge becomes obsolete very quickly.The most significant recent advances, in my view, are improved fidelity to references, character consistency, control over camera movement, image-to-video animation, and the ability to sustain an action for longer periods of time.
We are gradually moving from generation that feels almost like a lottery towards a tool that is beginning to allow for more precise direction. There is still an unpredictable element, but we can increasingly make more decisions ourselves and leave less to chance.
LBB> Where do you think AI is currently weak in the filmmaking process, where do you find that the classic methods are still best/easiest/quickest?
Pablo> AI is still particularly weak when it comes to portraying realistic, sustained human performance. But above all, AI is weak at representing emotion. And it is precisely in this anti-humanist aspect that the greatest distance remains between conventional productions and AI-generated ones.
Emotions are not merely an aesthetic expression. Reproducing sad eyes or a smile is not enough. An actor interprets what a character wants, what they are hiding, and the relationship they have with others. Within that sequence of decisions lies a truth that is impossible to obtain through artificial generation.
Beyond that, traditional methods also remain faster and more controllable when we need long dialogue scenes, complex physical actions, precise interaction between several characters, continuity across shots, or a completely faithful representation of a product.Sometimes, it is simply easier to place an object on a set and rotate it 45 degrees than to repeatedly explain to a model how it should do it.
LBB> One of the pushbacks we hear frequently is that AI filmmaking removes the fun that attracted many people to filmmaking in the first place – what’s your take? Where do you find the fun?
Pablo> Cinema, like any art form, is about expression and emotion. When I go into a cinema to watch a film, I expect to leave the theatre different from how I entered it. When I look at a painting, a sculpture, a building; when I listen to a song, read a book, or admire a piece of classical dance, I feel alive.
Is it possible to enter that humanist climax through AI? I am not the one to question how each person experiences emotion.
Personally, I find the fun in AI during the exploration phase. But I also think the fun disappears when the process is reduced to entering instructions and generating hundreds of images while waiting for something interesting to appear. At that point, we stop directing and start playing a visual slot machine.
That is why I am particularly interested in hybrid work. AI can expand the playing field, but people bring conversation, performance, collaboration, and creative accidents. The fun is not found in any particular tool, but in discovering the right way to express an idea.
LBB> The biggest challenge for major studios is legal uncertainty regarding IP and data. Recent legal cases (Getty v Stability) suggest that things may be flowing in favour of the tech platforms. Given there’s still no clear regulatory framework, what are the key legal and ethical concerns you prioritise in your production processes and how do you address them?
Pablo> For me, legislation needs to be closely connected to a society’s culture. The law must evolve alongside human beings and adapt to our ethical principles, our safety, and the protection of our rights.
Artificial intelligence is becoming increasingly and dangerously linked to misinformation and to the construction of a world in which anything goes: political manipulation, identity impersonation, the spread of fake images, ridicule, or the pursuit of a few likes on social media.
In the audiovisual field, I am particularly concerned about the unauthorised use of creative works, the generation of images or voices of people without their consent, data privacy, and the growing difficulty of distinguishing between real and manipulated content. As a society, we have a responsibility to confront these risks and demand rigorous, clear legislation capable of protecting us as creators, consumers, and citizens in an increasingly hyperconnected world. Technological development cannot advance at the expense of human rights, truth, or consent.
LBB> The digital ecosystem is filling with generic AI-generated content (or “AI Slop”). As a filmmaker, what is your creative philosophy for using AI to generate content that stands out and feels authentic, rather than plastic, idealized, or disposable?
Pablo> As artists – although the term itself is broad and abstract – we are always fighting against our own ego. The danger of generative AI is that it satisfies that part of us far too quickly. It is relatively easy to create, in a single day, a coherent piece that works reasonably well and is aesthetically correct. This brings us closer to a kind of content that is generic, immediate, easy, and ultimately tri
This is not limited to the audiovisual world. In everyday life, too, we run the risk of delegating an ever-greater part of our hands-on relationship with the world to AI, from thinking up a recipe to writing an email because the result seems grammatically better than what we would write ourselves. The problem is not using the tool, but ceasing to exercise certain abilities because receiving an immediate answer is more convenient.
All of this reassures the ego because it avoids error, and it happens in a matter of seconds. But I believe that what is human lies in error, authenticity, and imperfection. That is why I always try to move away from the excessively clean, perfect, idealised aesthetic that many models produce by default. Very often, what makes an image memorable is precisely the thing a model would try to correct.
LBB> Looking ahead 3-5 years, what is one creative constraint you have today (e.g., location, character consistency, budget scale) that you are certain AI will completely eliminate for commercial directors and producers?
Pablo> I think one of the major limitations that will disappear over the next few years is the difficulty of generating truly realistic movement, both in the camera and in characters. Today, AI still struggles to reproduce the weight, inertia, intention, and natural quality of human or cinematic movement with precision.
However, beyond purely creative advances, I would also like that evolution to be sustainable and fair. Progress should not be measured solely by what technology is capable of generating, but also by its environmental impact and by its respect for human, labour, and authors’ rights.
People in association withThe Monthly Cut
More News from SOMOS5 AUDIOVISUAL S.L.
For Julieta Vilaplana, Great Ideas Are Built Together22/07/2026
Growing Into Production with Alejandra Rodríguez01/07/2026
The Producer’s Job Today: Connecting Ideas, Technology and Execution with Diego Nonogawa Zimic19/05/2026
ALL THEIR NEWS
Work from SOMOS5 AUDIOVISUAL S.L.
Galaxy ZFold 8Samsung
14/08/2026
“El metro que soñamos”Metro de Madrid
27/07/2026
Sprite x Omar MontesSprite
27/07/2026
ALL THEIR WORK
SUBSCRIBE TO LBB’S newsletterEnter your email
FOLLOW US
Events
DPA + SLA (+ US Add.)
MyLBB Score
About Us
Contact
Editorial Toolkit
Help & FAQ
Privacy Policy
Main Terms & Conditions
Newsletters
Create your own LBB icons
How it works
Plans & Pricing
LBB’s Global Sponsor
