When the team behind Disco Elysium (2019) built Revachol, they didn't sketch a city — they built a coastline, a tidal basin, a waterfront that had once belonged to a naval empire. The water explained the docks, the docks explained the unions, the unions explained the politics, the politics explained the war. Every narrative thread ran back to geography.
Before cultures, economies, or magic systems exist, physical geography determines what is possible. Mountains create isolation and breed distinct languages. Rivers create trade, and trade creates conflict. Deserts limit population density; coasts encourage it. When geography is internally consistent, readers intuit the logic of a world even when they can't articulate why it feels real.
AI assistants — particularly large language models — have processed enormous volumes of geographic, geological, and ecological information. This makes them useful partners for pressure-testing invented geographies: Would this river actually flow that direction? What crops would grow here? What would people build with local materials?
In multiple public lectures at Brigham Young University (recorded 2020–2023), Sanderson described spending weeks researching how pre-industrial societies transported goods before writing The Way of Kings. The Shattered Plains — a fractured plateau landscape — emerged partly from thinking about how geography forces military strategy. AI tools now let writers run equivalent research queries in minutes rather than weeks.
Physical terrain doesn't just set the stage — it shapes culture. Anthropological and historical evidence is consistent: island civilizations develop navigation traditions, oral history transmitted by sea routes, and protein sources from the ocean. High-altitude cultures develop specific lung adaptations, terraced agriculture, and religious relationships with sky and storm. Forest civilizations build differently than steppe civilizations — their architecture, their gods, their warfare all differ.
When using AI for world-building, the terrain-culture link is your most valuable chain of inference. A single geographic prompt can cascade into dozens of cultural and narrative details.
Climate is geography's most direct storytelling tool. Seasons create natural pacing for narratives — medieval European literature is full of stories that begin in spring and resolve before winter. Monsoon cycles structure agriculture and military campaigns in South Asian epic traditions. When you establish a climate system with AI, you gain not just weather but a temporal architecture for your story.
A useful technique: describe your world's general latitude, ocean currents if any, and dominant terrain, then ask AI to suggest a realistic climate system and its implications for agriculture, seasonal travel, and warfare windows. The responses will often surface narrative possibilities you hadn't considered.
Start with one physical fact: "My world has a continent-spanning mountain range running north-south." Then cascade: What rain shadow effects exist? Where does desert form? Where do major rivers originate? Where do civilizations cluster? What do they fight over? Each answer constrains and enriches the next. AI is excellent at maintaining this cascade across a long conversation.
One of AI's most underused world-building functions is consistency auditing. Once you've described your world's geography in detail, you can ask the AI to identify logical contradictions: "Given this map description, would these two cities realistically have independent water supplies?" or "Is there a geographic reason this culture would have developed seafaring before their neighbors?" The AI acts as a first-pass geological and ecological editor.
Game designers at studios including Larian Studios (creators of Baldur's Gate 3) have discussed in GDC talks how geographic consistency creates the foundation for systemic gameplay — players internalize world logic and begin to predict how systems interact. The same principle applies to narrative fiction.
You're going to use the Cascade Prompt technique from Lesson 1. Start by describing one physical fact about a fictional world region — a terrain type, a body of water, a dominant feature. Then let the AI help you cascade outward into climate, resources, and implied culture.
Aim for at least 3 exchanges. The AI will guide you through the cascade.
J.R.R. Tolkien spent decades developing Quenya and Sindarin before The Lord of the Rings was published. His linguistic work — documented in the posthumously released History of Middle-earth series — was not decorative. Language shaped culture: Elvish phonology implied Elvish aesthetics, which implied Elvish architecture, which implied Elvish history. The depth readers sense in Middle-earth flows directly from this internal linguistic coherence.
The most common world-building mistake is treating culture as a list of exotic customs. Real cultures are systems — interlocking beliefs, practices, and structures where changing one element ripples through others. A culture's food practices relate to its agricultural base. Its agricultural base relates to its property law. Its property law relates to its kinship system. Its kinship system relates to its mythology.
AI can help you think systematically about culture in a way that's difficult to do alone. Because language models have processed enormous amounts of anthropological, sociological, and historical material, they can surface connections between cultural elements that a writer might not anticipate.
Le Guin — daughter of anthropologist Alfred Kroeber — described in her 1989 essay collection Dancing at the Edge of the World how anthropological thinking shaped her world-building. She treated invented societies as functioning wholes, asking what gender systems, kinship structures, and economic arrangements were implied by each other. Her 1969 novel The Left Hand of Darkness used a single biological difference (the elimination of fixed biological sex) to cascade through an entire culture's politics, mythology, and language. AI can replicate this cascade method on demand.
Names are readers' first contact with a culture's identity. Tolkien's method — building names from consistent phonological rules — created a sense of linguistic depth without readers needing to study linguistics. Modern writers can use AI to develop consistent naming conventions that imply cultural relationships.
Effective linguistic world-building prompts work from culture inward to sound: "This culture is matrilineal, lives in dense jungle, and practices ancestor veneration. What phonological tendencies and naming conventions might fit this profile?" The AI can suggest consonant patterns, syllable structures, and honorific systems that cohere with the cultural context.
Power structures are culture's skeleton. Feudalism, caste systems, council republics, theocracies, merchant oligarchies — each creates different patterns of conflict, alliance, and mobility that directly shape narrative. When you ask AI to develop a power structure, the most productive prompt strategy is to start from the economic base: who controls the scarce resource?
In 2022, game narrative designer Edwin McMillen discussed in an interview with Game Developer magazine how the social hierarchy of a world determines what kinds of stories are possible within it — specifically, which characters can plausibly act as agents of change. This is as true for novels as for games.
Mythology is a culture's attempt to explain its own physical and social world. If geography and social structure are well-developed, mythology almost writes itself — it will be the stories that justify the existing order, explain the terrain, and account for catastrophes. AI can help you work backward: given your world's geography and power structure, what mythological themes would a people develop?
The Elden Ring development team — documented in interviews with Hidetaka Miyazaki in Edge magazine (2022) — worked with author George R.R. Martin on a world mythology that was designed to be fragmentary and reconstructed through play. This reflects a sophisticated understanding: myths in real cultures are never complete, internally consistent documents. They accumulate contradictions. AI can help you build myth systems that have this quality of organic incompleteness.
The most vivid cultures grow from specific historical conflicts. Ask AI: "What cultural practices, taboos, and festivals would develop in a society that survived a catastrophic famine 200 years ago?" The specificity of the historical trauma generates specific cultural responses — not generic "hardy people" tropes but particular rituals, food-hoarding behaviors, social attitudes toward waste.
Choose one of these starting points, or invent your own: (A) a culture that controls the only freshwater source in a desert region, (B) a culture that survived a plague that killed 70% of its population three generations ago, or (C) a seafaring culture that has never seen a forest.
Work with the AI across at least 3 exchanges to develop naming conventions, social hierarchy, religious practices, and one major cultural taboo that emerge logically from your starting point.
Brandon Sanderson has formalized what he calls "Sanderson's Laws of Magic" — presented publicly in a 2013 blog post and discussed extensively in his BYU lectures. The first law: An author's ability to solve conflict with magic is directly proportional to how well the reader understands said magic. The implication: rules aren't constraints on creativity, they're what make magic narratively functional.
A system without limits has no drama. If a magic user can do anything, there's no suspense in whether they'll succeed. Rules create the conditions for failure, and failure is where narrative lives. The same principle applies to technology in science fiction: if faster-than-light travel is free and instant, there's no meaningful distance, no meaningful waiting, no meaningful separation. Constraints are the story's engine.
AI is valuable here precisely because it can help you find the implications of your rules before your readers do. A good magic system has internal logic that, when you follow it rigorously, generates plot problems you didn't anticipate — and often solutions you didn't plan. AI can run these implications forward on request.
The creators of Avatar: The Last Airbender — Michael DiMartino and Bryan Konietzko — have discussed in DVD commentaries and the 2010 Art of the Animated Series book how each bending discipline was grounded in a real martial arts style (Tai Chi for Waterbending, Hung Ga for Earthbending, Northern Shaolin for Firebending, Ba Gua for Airbending). This gave each style its own movement logic, which created genuine physical constraints on what each bender could do — which generated the visual language of conflict. Rules derived from real physical systems produce more internally consistent fictional systems.
The distinction between "hard" magic (fully defined rules, like Allomancy in Mistborn) and "soft" magic (mysterious, rule-light, like the Force in early Star Wars) is not a binary. It's a spectrum with narrative implications at every point. Hard magic enables problem-solving plots — the reader can follow the logic and anticipate solutions. Soft magic enables awe and mystery — the unknown is the point.
When using AI to develop a magic system, clarifying your position on this spectrum is the first essential prompt. "I want a magic system where readers can understand the rules well enough to feel clever when characters find solutions" leads to very different AI output than "I want magic that feels ancient and unknowable to characters and readers alike."
Science fiction writers face a version of the same challenge: technology that solves all problems is narratively inert. The most compelling SF worldbuilding maintains technological scarcity and limitation even in advanced settings. William Gibson's cyberpunk — developed across the Sprawl trilogy beginning with Neuromancer (1984) — is technically advanced but brutally constrained by economic and social scarcity. Technology is plentiful; its distribution is not.
AI can help you identify anachronisms — technologies that shouldn't exist given your world's development state — and can also help you think through second-order effects: "If my civilization has antigravity but not germ theory, what does their medicine look like? What does their warfare look like?" These second-order questions generate the texture of a believable future or alternate history.
Once you've established a rules system with AI, run adversarial prompts against it: "Given these rules, what would a clever villain do to exploit the system?" and "What is the single most powerful ability this system allows, and does that break my story?" AI is particularly good at finding the logical extremes of a rule system — which is exactly where plot holes live.
Across game design, fantasy fiction, and science fiction world-building, a consistent principle emerges: every significant power needs a cost. The cost doesn't have to be equal — but it must be real and felt by the character. Magic that costs nothing is a deus ex machina waiting to happen. Magic that costs something the character values creates genuine choices, which create genuine drama.
When developing costs with AI, specificity matters. "Using this magic costs mana" is weak — mana is an abstraction. "Using this magic accelerates the user's aging by one year per use, and the effect is visible immediately to observers" is specific, creates social consequences, creates personal stakes, and generates narrative possibilities the author didn't have to invent alone.
Design a fictional rules system — magic, technology, or ability — in collaboration with the AI. Start by positioning yourself on the hard/soft spectrum, then develop the system's core mechanism and cost. After at least 2 exchanges of design, ask the AI to stress-test it by finding exploits and potential plot holes.
Complete at least 3 total exchanges to finish the lab.
When Kim Stanley Robinson developed the terraforming economy of Mars across his Mars Trilogy (1992–1996), he grounded conflict not in personal animosity but in competing economic models: corporate extraction vs. ecological preservation vs. Martian sovereignty. The conflicts felt real because they emerged from genuine resource scarcity and ideological disagreement about who a world belongs to. The economics preceded the drama.
Most genre fiction treats economy as background texture — the marketplace the characters walk through. But economy is the structural determinant of almost every major conflict. Wars happen over resources, trade routes, or the disruption of existing arrangements. Revolutions happen when economic mobility is blocked and resentment accumulates. Love stories are complicated by economic disparity. Even personal feuds often have economic roots that predated the interpersonal conflict.
AI can help you build economies that generate story rather than just providing backdrop. The key insight: what is scarce, who controls scarcity, and who is excluded from that control — answer these three questions and your major conflicts write themselves.
Frank Herbert's Dune (1965) is frequently cited in academic discussions of political ecology — the study of how resource control shapes political systems. The spice melange powers interstellar travel, extends life, and enables prescience. Herbert's genius was making this single substance simultaneously the foundation of the Imperium's power structure, the cause of the novel's central conflict, and the source of Arrakis's ecological importance. In his notes (published in 2019's The Road to Dune), Herbert described deliberately modeling the spice on oil — a single resource whose control determines the shape of an entire civilization. This is the Economy-First approach at its most sophisticated.
You don't need to be an economist to build a compelling fictional economy — but you do need to think about flow. Who produces? Who distributes? Who consumes? Where does the value accumulate? What happens to people who are excluded from the value chain? These questions, run through an AI conversation, generate the texture of daily life for every social class in your world.
Game worlds have driven some of the most sophisticated fictional economic thinking in recent years. Eve Online's player-driven economy — documented extensively by CCP Games' in-house economist Eyjolfur Gudmundsson between 2007 and 2014 — demonstrated that fictional economies with genuine scarcity and player agency produce authentic economic behavior including monopoly formation, arbitrage, and economic warfare. The lesson for narrative world-building: scarcity and consequence are the engines of economic story.
A world's history is not its backstory — it's the pressure bearing on the present. Characters in well-built worlds feel the weight of historical decisions they didn't make. The children of a conquered people carry the conquest in their language, their names, their religion, their relationship to law. The heirs of an empire carry guilt or denial. History that only exists as exposition is dead; history that lives in characters' bodies and choices is alive.
The technique for generating living history with AI is to focus on what was never resolved. Real historical conflicts don't end — they transform. The dissolution of the Soviet Union in 1991 didn't resolve the tensions it had suppressed; it released them into new forms. Ask AI: "Given this world's 200-year history, what conflicts from that period are still unresolved, transformed, and operating underground in the present day?"
Build your world's history as a set of unresolved debts rather than settled events. Ask AI: "What promises were made and broken in this world's founding? Who remembers, who has forgotten, and who profits from forgetting?" This generates both the texture of the present and the seeds of your story's central conflicts.
Military conflict in well-built worlds is always downstream of economic or resource conflict — even when the characters believe they're fighting for honor, religion, or national pride. The Peloponnesian War had trade route economics beneath its surface. The Crusades had Italian city-state commercial interests beneath their religious framing. Understanding this doesn't make your fictional conflict cynical — it makes it specific. Specific conflicts have shape, have legitimate grievances on multiple sides, and resist easy moral resolution. That complexity is what makes them narratively rich.
When building conflict with AI, the most generative approach is to ask for the economic interests of each party before establishing their stated motivations: "What does each faction economically gain or lose from this war?" The gap between stated motivation and economic interest is often where your most interesting character work lives.
You're going to build the economic spine of a fictional world and its most significant historical conflict — one that isn't fully resolved in the present. Start with a scarce resource and who controls it, then work with the AI to develop the factions it creates, the historical conflict that resulted, and the unresolved tensions that persist today.
Complete at least 3 exchanges to finish the lab.