The debate about AI is slowly moving away from the question of whether AI should be used and towards a more nuanced one: how, when, and why generative AI could be used. While much of that conversation focuses on image generation and productivity, some creators are using AI to tackle something far bigger: building entire worlds. NUVAGAIA is one of those projects.
Created by David Triguero, NUVAGAIA is an ambitious sci-fi universe that blends solarpunk ideals with cosmic horror and spans podcasts, stories, artwork, videos, a growing wiki, and even interactive fiction. It's the sort of expansive transmedia project that, not so long ago, would have needed a team of writers, artists and producers to get off the ground. It would have cost millions of dollars, teams across the globe, and most likely have been bogged down in paperwork and creative wrangling; just take a look at Reflector Entertainment’s Unknown 9: Awakening, which, while fun and polished, was in hock to its wider media ambitions.
Instead, Triguero is building this world largely himself, using a mix of AI tools to help develop lore, create visuals, generate voices and keep an ever-expanding canon on track. What's interesting isn't the technology itself, but how it's being directed, edited and shaped into something coherent.
Below, Triguero shares insights into how and why he’s creating a transmedia world with AI, why consistency matters more than spectacle, and how emerging tools are making it possible for solo creators to think on a much bigger scale than ever before.
CB: Why did AI become central to building NUVAGAIA as a transmedia world?
David Triguero: AI became central because NUVAGAIA began as an idea that was too large for me to develop alone through traditional means.
The first step was to create my own sci-fi world. But I soon understood that a world, by itself, does not reach people unless it finds ways to be discovered. That is where transmedia came in: podcast, stories, images, videos, website, and public lore as different entry points into the same universe.
AI allowed me to turn that vision into something producible by a single person. Under my creative direction, AI makes it possible to explore, prototype, and publish at a scale that would otherwise have been unfeasible.
CB: What does the design and world-building process look like before you use AI?
DT: In my case, there was no phase “before AI” that was completely separate. AI was there from the beginning, including research and conceptual exploration.
The first step was a strategic search: I wanted to create a science fiction IP, and I needed to understand which territories had potential and were still underdeveloped. That is how I discovered solarpunk, which I did not know by that name, and I began building from a premise: solarpunk in contrast with cosmic horror.
From there, I started creating documents, visual ideas, internal rules, civilisations, characters, and conflicts. AI helped me research faster, organise references, explore possibilities, and turn scattered intuitions into a clearer world structure.
CB: How are you using AI beyond writing or generating dialogue?
DT: I use it across many parts of the process, always as a directed tool. For images, I start from structured prompts, test variations, discard many outputs, and keep iterating until I find something that fits the visual tone of NUVAGAIA.
For the podcast, after working on the script, I use text-to-speech tools to generate the voices, then I manually handle the assembly, editing, music selection, effects, and pacing. I also use it for the website as a vibe-coding process. In addition, it helps me organise lore, test ideas, maintain continuity, translate, prepare communication materials, and test how the same idea works across different formats.
CB: How do you make an AI-assisted world feel believable and coherent across different formats?
DT: Coherence does not appear simply because you use AI; it appears because you have clear criteria. NUVAGAIA has internal rules, names, civilisations, a chronology, an aesthetic, and its own logic for its conflicts. AI can help expand all of that, but it does not decide what belongs in the world and what does not. That decision is mine. If an image is spectacular but does not feel like NUVAGAIA, it is not useful. If a text explains too much or contradicts the tone, it gets rewritten.
I try to make each format serve a function: the podcast brings immersion, the stories deepen the characters and culture, the wiki organises the canon, and the images make the world visible. AI accelerates the process, but coherence comes from my vision and judgment.
CB: Can NUVAGAIA evolve beyond a fixed story?
DT: Yes. NUVAGAIA is not conceived only as a closed story, but as a universe that can grow in layers. Right now, it unfolds through podcasts, stories, a wiki, images, and videos, but I am already testing more open forms.
The first example is the Thresholds series: interactive pieces where readers can decide where the story goes next. That choice does not remain external or anecdotal; it becomes part of the canon and is incorporated into the wiki. In the long term, NUVAGAIA could expand into other formats like tabletop or card games, or evolve toward a more interactive experience, a video game, a series, or a film.
But before reaching that point, the important thing is to build a solid foundation: world, tone, characters, rules, and community.
CB: How do you maintain continuity of characters, civilizations, and events across different formats?
DT: I build NUVAGAIA from the inside out. There is a lot of internal material that the public may never see, but which supports what gets published: character sheets, civilisation guides, names, chronology, rules for the Star Weavers, visual tone, and relationships between events.
The podcast tells the main story, but the short stories allow me to expand on what does not fit there: character backstories, everyday life, cultural tensions, or small stories from the world. The wiki functions as a public archive of the canon, which is progressively unlocked and expanded.
AI serves as a working memory and consultation tool, but continuity depends, above all, on structure, revision, and human editorial decisions.
CB: What AI models do you use and why?
DT: I use different tools depending on the part of the process. ChatGPT is my main assistant for worldbuilding, organisation, creative strategy, and coherence review, because it allows me to work in a structured way and maintain a global view of the project. Gemini works well for supporting the development of podcast scripts; after testing several options, it is the one that works best for me.
For voices, I use Gemini TTS within a small customised environment in Google Colab, because it allows me to control tone, speed, and intention. For images, I have mainly used DALL·E 3 because, although it predates more recent models, it gives me the hand-drawn finish I am looking for in NUVAGAIA’s visual identity. I also use ComfyUI, GPT Image 2, Nano Banana Pro, and Veo in Google Flow for retouching, editing, and video.
CB: How do you stop AI-generated content from feeling repetitive or soulless?
DT: You should not accept the first output as the final answer. AI can produce correct things very quickly, but the correct does not always have soul. That is why I work with clear limits: which civilisation is speaking, what tone it has, what conflict lies underneath, and what function that piece serves within the world. Many ideas are discarded. Others are transformed heavily before being published.
In images, I look for expressive imperfection and coherence, not just spectacularity. In texts, I try to make sure there is always a human question underneath: loss, identity, hope, guilt, community. Where I see the greatest difficulty is in AI voices. They have advanced a lot, but they can still break immersion. That is why I treat the current podcast as an evolving proof of concept.
CB: What inspired your approach to AI-powered world-building?
DT: I had previously worked on another personal AI-assisted project, and from that moment I saw the potential and decided to try to build my own sci-fi IP. What interested me first was the world: its cultures, its aesthetic, its tensions, and its possibilities. The initial research led me to solarpunk, and when I crossed it with cosmic horror, the tension that gave birth to NUVAGAIA appeared: a hopeful future confronted with something it cannot fully understand.
Later, I realised I needed more variety and developed the Hydrovelan and Nanocodax civilisations as my own evolutions of steampunk- and cyberpunk-like imaginaries. AI allowed me to explore all of that at great speed, but the inspiration was not purely technological; it was creative: using a new tool to explore an idea that, without it, would probably have remained in a drawer.
CB: Do you see AI as a production tool or a new way to tell transmedia stories?
DT: Right now, I use it mainly as a tool for production and exploration, but I believe it can open up new ways of telling transmedia stories. In NUVAGAIA, it allows me to move from one idea to multiple expressions: story, image, episode, video, website, wiki, or interactive piece. That changes the way you think about a fictional world, because you do not develop it only as text, but as a system of connected formats. Even so, I do not believe AI is an author by itself. For me, authorship lies in vision, judgment, selection, and direction. AI is a new palette of brushes: it expands what you can attempt, but it does not decide what deserves to exist.
CB: Could you do this project without AI?
DT: NUVAGAIA as an idea, perhaps yes. NUVAGAIA as a real, published and transmedia project, not in this way. Without AI, it would probably have ended up as a folder of notes, a few concepts, and maybe a short story. AI greatly reduced the distance between imagining the world and being able to materialise it. In one week, I was able to lay down worldbuilding foundations that would have taken me months or years on my own. After that, it allowed me to develop images, scripts, stories, videos, and the website. That does not mean everything is created by pressing a button. There is direction, editing, and a lot of manual work. But it does mean that one person can build a modest project that previously would have required an entire team.
CB: What do you say to people who have yet to use AI?
DT: I would tell them to start with their own vision, not with the tool. If you do not have judgment, AI can easily lead you toward the generic. But if you know what you want to say, what tone you are looking for, and what boundaries you do not want to cross, it can empower you and become an incredible creative workshop. I also understand the doubts. There are legitimate concerns about ethics, rights, the replacement of human work, and the loss of artistic value. I do not think those concerns should be ignored. But in my experience, using AI does not eliminate human creativity; it amplifies it. The more powerful the tool, the more you need taste, direction, editing, and responsibility. For me, it is not about letting AI create for you, but about learning how to direct it.
