STRATEGYOngoing practice88% confidence

The Fiction-Reality Gap

Shared stories drive cooperation — divergence from reality breeds conflict

Problem it solves

Why wars and crises emerge without requiring bad actors

Best for

Understanding why macro conflicts, market bubbles, and political ruptures happen; framing crypto legitimacy; assessing AI misinformation risk at structural scale

Not ideal for

Near-term trade timing or short-term price action; the framework operates at civilizational timescales

Overview

Why this framework exists

Harari's central argument is that humans dominate the planet not through individual intelligence but through the ability to cooperate flexibly in massive numbers via shared fiction. Money, nations, corporations, and religions are collective stories that exist only in belief — and when enough people believe them, they function. The danger is not fiction itself but the widening gap between those fictions and physical reality.

AI compounds this risk categorically. Unlike the printing press or radio, which disseminated human-generated content, AI generates novel content and makes independent decisions. When algorithms optimize for engagement rather than truth — as social media already does — they accelerate fiction-reality divergence. Generative AI adds a new layer: fictions produced at scale with no human authorship, making the gap harder to detect and govern.

The financial system serves as the clearest preview. CDO complexity in the 2007–08 crisis showed how instruments too complex for any regulator to understand could destabilize the global economy. AI-generated financial instruments could repeat this at orders of magnitude greater complexity, with the additional property that no human understands them even in retrospect — removing political accountability entirely.

Core principles

5 total
  1. All human large-scale cooperation runs on shared fictions — money, nations, corporations, religions — that have no objective existence beyond collective belief
  2. A fiction's danger is not that it is false but that it diverges from physical reality faster than society can adapt its behaviour
  3. AI is the first technology that both generates fictions autonomously and makes consequential decisions, making it categorically different from all prior media technologies
  4. When algorithms optimize for engagement rather than truth, fiction-reality divergence becomes structurally self-reinforcing
  5. Financial complexity is the canary: instruments no human can understand are the prototype for AI-generated fictions that destabilise systems no political body can govern

Steps

5 steps
  1. Map the operative fictions in your domain
    Identify the shared stories that coordinate behaviour in your area — asset valuations, regulatory legitimacy, brand trust, national narratives. These are the things people act on that have no physical existence beyond collective belief.
    Pro tipCrypto assets are a clean test case: Bitcoin's value is explicitly a collectively maintained fiction, which is also how all fiat money works. Recognising this is not a dismissal but an accurate description of mechanism.
  2. Assess the fiction-reality gap
    For each operative fiction, estimate how far it has drifted from verifiable physical reality. The gap itself is not automatically dangerous — stability depends on whether the drift is shared and stable, or accelerating and contested.
    WarningRapidly widening gaps — where the fiction becomes detectable as fiction to a significant minority — are the leading indicator of rupture, not the gap size in absolute terms.
  3. Identify the fiction-generation mechanisms
    Determine what produces and maintains the operative fictions: institutions, media ecosystems, algorithmic curation, or generative AI. The mechanism determines how quickly fictions can shift and how detectable shifts are.
  4. Evaluate feedback loops and speed asymmetries
    Assess whether the feedback loop between fiction-reality divergence and corrective action is intact. AI-generated financial instruments and deepfakes both break this loop by making divergence undetectable until after catastrophic consequences.
    WarningSpeed asymmetry is the critical variable: if fictions can shift faster than institutions can adapt, stability collapses even without bad-actor intent.
  5. Stress-test your exposure to fiction collapse
    For any position, narrative, or strategy that depends on a shared fiction holding, model what happens if 20%, then 40%, then 60% of participants withdraw belief. The initial shock is usually the smallest effect — cascading disbelief is the systemic risk.
    Pro tipApply this to sovereign debt, crypto assets, brand valuations, and regulatory legitimacy simultaneously to identify correlated exposures.

Checklist

Saved in your browser

Examples

3 cases
Bitcoin as a functional fiction

Harari explicitly uses Bitcoin as an example of collective fiction that functions precisely because the mechanism is the same as fiat money. 'You cannot eat them or drink them. But you have people telling very compelling stories about the value of these things and if enough people believe the story then it works.' This is not a critique of Bitcoin but an accurate description of how all monetary value operates.

OutcomeRecognising crypto as a collectively maintained fiction reframes the key risk question: not whether it has 'intrinsic value' but whether the narrative infrastructure maintaining collective belief is more or less robust than competing monetary fictions.
2007–08 CDOs as AI prototype

A small number of people created financial instruments of sufficient complexity that regulators, investors, and governments could not interrogate them. The fiction — that these instruments were safe, rated assets — diverged from reality, and the collapse was systemic. Harari argues AI-generated financial instruments are the same failure mode at higher complexity and speed, with the additional property that even the creators may not understand them.

OutcomeThe CDO crisis provides an empirical baseline for what AI-generated financial complexity risk looks like. Harari's projection: 'Fast forward 10 or 20 years. AI creates such complicated financial devices that there is not a single human being on earth that understands finance anymore.'
War over mythologies, not resources

Harari uses the Israel-Palestine conflict as a case where the physical resources (land, food, water) are sufficient for all parties but the fiction gap — different mythologies about who the land belongs to, different national stories — makes coordination impossible. The conflict is not about scarcity but about incompatible shared stories.

OutcomeThis reframes geopolitical risk assessment: resource adequacy does not predict peace. The operative variable is shared narrative infrastructure, which is what AI-generated misinformation directly attacks.

Common mistakes

4 traps
Treating fiction as synonymous with falsehood
The framework explicitly holds that shared fictions can be entirely functional and stability-producing. The error is assuming a fiction's falsity makes it fragile; what matters is whether the collective belief is stable, not whether the underlying claim is physically true.
Underestimating the category shift from AI content generation
Most analyses treat AI misinformation as a scale problem — more of the same kind of thing social media already produced. Harari's point is that AI content generation is a category change: fictions with no human author cannot be traced, attributed, or governed through existing accountability structures.
Applying the framework only to political narratives
The Fiction-Reality Gap is most commonly discussed in political contexts. Its equal application to financial instruments (CDOs, AI-generated derivatives), brand trust, and regulatory legitimacy is systematically underused — and these domains have clearer near-term risk implications.
Assuming stable fictions are permanent
The historical record shows that extremely stable fictions — the divine right of kings, gold's monetary status — can collapse rapidly once a significant minority begins to defect. Stability is not evidence of permanence.

Origin story

How this framework came to be

This framework is the core thesis of Harari's 2011 book Sapiens: A Brief History of Humankind, developed from his work as a historian at Hebrew University of Jerusalem. The key insight — that Homo sapiens' competitive advantage over other species was the ability to cooperate around shared myths — emerged from his study of the cognitive revolution approximately 70,000 years ago.

By his 2015 book Homo Deus and subsequent work, Harari extended the framework to cover the specific vulnerabilities that AI and algorithmic media introduce: not merely the dissemination of fictions, but their automated generation at scale, with optimization functions that have nothing to do with social stability.

Source

Traced to primary
Source · PODCAST
Yuval Noah Harari: An Urgent Warning They Hope You Ignore. More War Is Coming!
Yuval Noah Harari · 2023
Open source →

Related frameworks

Browse all Strategy →