1
A fantasy writer supplies five fixed magic rules and asks for four second-order effects on emergency medicine, then manually chooses one and updates the world bible.
16 AI Worldbuilding Examples
Use these examples to study structure, specificity, tone, and variation. They are demonstrations—not claims about a real person or organization unless the example itself makes that explicit.
Before you copy
What to notice in the examples
A strong AI worldbuilding workflow starts from author-supplied canon and constraints, asks for multiple options rather than one authoritative answer, labels inventions and research needs, verifies real-world claims independently, checks privacy and contractual limits before uploading material, and records only deliberately accepted ideas as canon. The writer remains responsible for originality, consistency and the delivered work.
- Canon packet: provide only the fixed facts and constraints needed for the current task.
- Narrow job: ask for alternatives, consequences, contradiction detection, research questions or comparison—not “build my world.”
- Output labels: separate invented possibilities, assumptions, uncertain claims and items that require external verification.
- Human selection: choose, combine, reject and rewrite based on story purpose and established canon.
- Verification: independently check factual, cultural, historical, scientific and legal claims before relying on them.
- Privacy/rights gate: do not upload confidential manuscripts, private research or third-party text when the tool’s terms or permissions make that unsafe.
- Canon update: record only accepted decisions in the human-controlled series/world bible, with provenance where useful.
2
A science-fiction writer gives a fixed communication delay and asks for three contract systems that could plausibly emerge, labeling any real legal analogy as research-needed.
3
A series author provides a short canon list and asks the model to flag contradictions created by a proposed political alliance, then checks every flag against the actual manuscript notes.
4
A mystery writer asks for six fictional transit-ticket mechanisms that could create alibis, but independently researches any real payment-system assumptions before adapting one.
5
A historical-fantasy writer asks AI to generate research questions about textile production rather than requesting invented period facts or citations.
6
A writer asks for four cultures responding differently to the same water shortage, then rejects options that collapse culture into climate determinism.
7
An author uses AI to compare three inheritance systems against supplied character goals, choosing the one that creates the most useful conflict rather than the most elaborate lore.
8
A writer provides a canon timeline and asks for impossible overlaps, then verifies flagged dates manually before changing the series bible.
9
AI proposes a ritual; the writer keeps only its structural function, researches real-world analogues carefully, and rewrites the practice to fit the invented belief system.
10
A worldbuilder asks for ordinary inconveniences created by teleportation, then traces one selected inconvenience into housing, insurance and policing without copying the generated prose.
11
A writer with confidential interview notes summarizes only the non-sensitive constraint instead of uploading the transcript to a consumer chatbot.
12
A novelist asks for five failure modes of a fictional energy grid and requires each answer to name the assumption it depends on.
13
A writer asks AI to find gaps in a fictional economy, then treats numerical claims as placeholders until independently researched or removed.
14
A project uses AI only to generate alternative questions for a culture audit; the author answers those questions from their own canon and research.
15
A writer detects a suspiciously specific phrase in generated text, discards it, and retains only the underlying independently developed idea.
16
Before delivery, an author reviews publisher contract terms and copyright-registration implications of any retained AI-generated material rather than assuming tool use is invisible or irrelevant.
Keep the underlying decision or pattern, then replace the subject, evidence, relationship, constraints, and tone with details that belong to your situation. If your final line still works after swapping only one noun, it may be too close to the example.