The Collapse of Shared Reality
The shared factual basis for collective decision-making erodes.
A megatrend's gradient is the minimum maturity of its enablers, not the mean.
Epistemic Fragmentation
shared factual floor fragments into competing epistemic communities
| Enabler | Family | Weight | Maturity | ||
|---|---|---|---|---|---|
| Epistemic Trust | human | large | 2/5 | ○ candidate constraint | |
| Social Credit System | technological | marginal | 4/5 | marginal | marginal: not counted in the minimum |
| Generative Media | technological | large | 5/5 | ||
| Large Language Models | technological | moderate | 5/5 | ||
| Recommender Systems | technological | large | 5/5 | ||
| Social Networks | technological | moderate | 5/5 |
Clock reading
- Individuals — months: Adapted sideways: belief formation moved to peer networks rather than improving.
- Companies — years: Platforms profit from the fragmentation; everyone downstream decides on the fragmented information.
- Governments — decades: Slow, and undermined by the trust decline that runs alongside it.
- The biosphere — not ours to set: Coupled through collective-action failure.
What would prove this wrong
Provenance standards deployed across capture, distribution and consumption simultaneously — high conjunction, and therefore unlikely soon.
What this pushes on
- amplifies → Polarisation and Radicalisation: Separate factual worlds make disagreement irreconcilable rather than negotiable.
- amplifies → Institutional Trust Decline: Institutions cannot be trusted by people who do not share their evidence.
- amplifies → The Diagnostics Gap: Contested measurement is measurement that cannot ground a decision.
Polarisation and Radicalisation
populations sort into communities that hold incompatible pictures of the same events
| Enabler | Family | Weight | Maturity | ||
|---|---|---|---|---|---|
| Generational Turnover | human | moderate | 2/5 | ○ candidate constraint | |
| Generative Media | technological | moderate | 5/5 | ||
| Recommender Systems | technological | moderate | 5/5 | ||
| Social Networks | technological | moderate | 5/5 |
Clock reading
- Individuals — months: Fast, and asymmetric — a position hardens faster than it softens.
- Companies — years: Platform firms are upstream of it and monetise the engagement it produces; everyone else absorbs the cost.
- Governments — decades: Slow, and compromised: a polarised legislature is the institution least able to act on polarisation.
- The biosphere — not ours to set: Coupled through the collective-action failure it produces on every environmental question.
What would prove this wrong
Affective polarisation falling across several democracies simultaneously, or a measured recovery in cross-group trust without a unifying external shock causing it.
What this pushes on
- dampens → Government as a Service: Polarised legislatures do not fund unglamorous service reform.
Institutional Trust Decline
public confidence in mediating institutions erodes across democracies
| Enabler | Family | Weight | Maturity | ||
|---|---|---|---|---|---|
| Epistemic Trust | human | large | 2/5 | ○ candidate constraint | |
| Cryptography & Privacy Technologies | technological | moderate | 3/5 | ||
| Cybersecurity | technological | moderate | 4/5 | ||
| Social Credit System | technological | moderate | 4/5 | ||
| Generative Media | technological | moderate | 5/5 |
Clock reading
- Individuals — months: Fast to lose, slow to rebuild, and the asymmetry is the mechanism.
- Companies — years: Not the actor, though the vacuum is commercially exploitable.
- Governments — decades: Decades to rebuild, and each failed programme resets it.
- The biosphere — not ours to set: Coupled — a low-trust polity cannot execute a transition that requires sustained sacrifice.
What would prove this wrong
A sustained recovery in institutional trust in an OECD country without an external shock producing it — driven by a record of delivery rather than by a rally effect.
What this pushes on
- amplifies → Climate Resilience Imperative: Long-horizon collective action requires trust the system no longer has.
- amplifies → Fertility Collapse: ⚠ Associational only, with one quasi-experimental study. Not a clean causal claim.
Synthetic Media and the Verification Collapse
cost of producing convincing synthetic content falls toward zero
| Enabler | Family | Weight | Maturity | ||
|---|---|---|---|---|---|
| Regulatory Clarity | structural | large | 2/5 | ○ candidate constraint | |
| Digital Identity & Verification | technological | large | 3/5 | ||
| AI & Machine Learning | technological | large | 5/5 | ||
| Generative Media | technological | large | 5/5 |
Clock reading
- Individuals — months: Already adapted, by lowering trust in all media rather than by learning to distinguish.
- Companies — years: Detection tools exist; the scaling problem is not solved by any of them.
- Governments — decades: Liability and provenance rules are early everywhere and settled nowhere.
- The biosphere — not ours to set: Not coupled.
What would prove this wrong
Authentication standards deployed widely enough that the liar's dividend shrinks — that is, that denying authentic media stops working.
What this pushes on
- amplifies → Epistemic Fragmentation: When anything can be fabricated, anything real can be dismissed.
- amplifies → Institutional Trust Decline: The liar's dividend applies to institutional evidence first.
The Diagnostics Gap
the capacity to produce information outruns the capacity to evaluate it, at every scale from the citizen to the state
| Enabler | Family | Weight | Maturity | ||
|---|---|---|---|---|---|
| Epistemic Trust | human | large | 2/5 | ○ candidate constraint | |
| Sensors & Earth Observation | technological | large | 3/5 | ||
| Data & Datafication | technological | large | 4/5 estimate | ||
| AI & Machine Learning | technological | moderate | 5/5 | ||
| Large Language Models | technological | moderate | 5/5 |
Clock reading
- Individuals — months: Adapted sideways rather than upward: belief formation migrated to peer networks rather than improving.
- Companies — years: Downstream consumers of a broken epistemic environment, deciding on the same fragmented information their regulators use.
- Governments — decades: Slowest, and worst placed — a government that has lost trust cannot run an information-integrity programme without it being received as propaganda.
- The biosphere — not ours to set: Coupled through the inability to agree that a problem exists, which precedes every response to it.
What would prove this wrong
Verification infrastructure deploying at the scale of production — provenance standards adopted across capture devices, platforms and regulators simultaneously — or education systems teaching diagnosis rather than detection at national scale.
What this pushes on
- amplifies → The Application Gap: Organisations that cannot see their own state fund the visible layer.
- amplifies → Government as a Service: A state that cannot measure its own delivery cannot improve it.
What holds this force back
- Epistemic Trust
Sources for this force
- C2PA specification status; NASEM and platform policy record (2025–2026)
- Contested — large-scale field experiments vs observational literature; see the register's social-media rows (2020–2025)
- DOCUMENTED ABSENCE, re-confirmed Sep 2026. Nearest institutional evidence: OECD Going Digital Measurement Roadmap (2026); OECD Digital Government Outlook (2025/2026); Data Governance Hub 4th Annual Report (2024). No primary measurement of data fitness exists (2024-2026)
- E-052 (adoption); E-104 (price); US Census BTOS (2023–2026)
- E-168 (BLS ATUS); E-075 (GSS via IFS); contested effects literature (2024–2025)
- eIDAS 2.0 implementation timeline; Aadhaar and Estonian deployment record (2024–2026)
- ENISA Threat Landscape 2025 (Oct 2025, ref period Jul 2024–Jun 2025): 4,875 curated incidents; DDoS ~77% of reported incidents, phishing ~60% and vulnerability exploitation ~21.3% as leading vectors. CISA/ICS-CERT: 3,637 ICS advisories covering 12,174 vulnerabilities across 2,783 products (Mar 2010–Jan 2026). CISA Strategic Intent (2021): ICS 'often fail to receive software patches in a timely manner.' ISACA State of Cybersecurity 2025: 55% understaffed, 65% with unfilled positions (2024–2026)
- GDPR 2012→2018 and EU AI Act 2021→2024→2027 timelines; Marchant, Allenby & Herkert (2011) (2011–2027)
- MethaneSAT (launched Mar 2024; $88m, $100m Bezos Earth Fund grant, NZ$32m New Zealand) lost contact 20 Jun 2025 and was declared unrecoverable 15 months into a five-year mission — having measured Permian Basin emissions at 3–5× EPA estimates. ESA Sentinel-5 provides broad-scale coverage (2024–2025)
- NIST post-quantum standards and IR 8547 (E-176); deployment surveys (2024–2026)
- OECD Survey on Drivers of Trust in Public Institutions (2021, 2023, 2024); Eurobarometer; V-Dem (E-046) (2024–2026)
- Stanford HAI AI Index; provider pricing series (E-104); US Census BTOS (2025–2026)
- UN World Population Prospects 2024 (E-001, E-003, E-005) (2024)
- WKO Innovation Map 26/27 — TRL 8 / SRL 3, a +5 gap (E-142) (2026)