LIVING SYSTEMS INTELLIGENCE
Intelligence integrity

Trust

A serious intelligence system is open about what it knows and what it does not. This page tracks the evidence base — claims, sources, confidence, review status and the gaps still to be closed.

31
Claims
16
Sources
4
Need sourcing
10
Sources to verify
Solution & decision intelligence integrity
8
Solution pathways
6
Decision signals
1
Signals to source
1
Pathways to evidence

Decision signals are structured reasoning, not automated advice. Items listed above still need stronger sourcing or evidence before they should be relied upon.

Amazon — flagship case coverage
15
Amazon claims
7
Amazon sources
4
Need sourcing
10
Data gaps
The Amazon rainforest stores large amounts of carbon, supporting climate regulation.Needs source
The Amazon rainforest provides habitat for a very large share of terrestrial biodiversity.Needs source
Weakening rainfall regulation may create risks for agriculture and water systems.Needs source
Amazon ecosystem degradation can create cascading risks across ecological and human systems.Draft
Claims by confidence
High6
Medium23
Low2
Uncertain0
Claims by review status
Verified5
Reviewed22
Needs source3
Draft1
Nodes needing stronger sourcing (4)
The Amazon rainforest stores large amounts of carbon, supporting climate regulation.Amazon RainforestNeeds source
The Amazon rainforest provides habitat for a very large share of terrestrial biodiversity.Amazon RainforestNeeds source
Weakening rainfall regulation may create risks for agriculture and water systems.Rainfall RegulationNeeds source
Amazon ecosystem degradation can create cascading risks across ecological and human systems.Amazon RainforestDraft
Relationships needing stronger evidence (4)
The food system depends in part on pollination through crop production.DEP_FOOD_POLLINATION
Pollinator decline can weaken the pollination that crops rely on.DEP_FOOD_POLLINATION
Rainfall regulation supports regional agriculture and water availability.DEP_AGRI_RAINFALL
Weakening rainfall regulation may create risks for agriculture and water systems.DEP_AGRI_RAINFALL
Known data gaps (18)
  • — Estimates vary widely between regions and surveys.
  • — Relative contribution varies by crop and region.
  • — Forage availability not yet mapped by region.
  • — Relative weight of each driver remains uncertain.
  • — Exact share depends on crop mix and measurement method.
  • — Quantitative dependency strength not yet modelled.
  • — Field and laboratory effect sizes differ.
  • — Specific carbon-stock sources still to be itemised.
  • — Strength of the effect varies regionally.
  • — Quantitative dependency strength not yet modelled.
  • — Fire, climate and deforestation interactions are complex.
  • — Outcomes vary by territory and governance context.
  • — Extent and mercury impacts not fully mapped.
  • — Outcomes depend on method, scale and time.
  • — Attribution by commodity is partial.
  • — Thresholds and regional variation not yet quantified.
  • — Cascade magnitudes are illustrative, not yet quantified.
  • — Outcomes depend on practice and context.
Learning intelligence integrity
6
Expected outcomes
6
Observed outcomes
6
Learning records
2
Confidence updates
By assumption status
  • PartlySupported3
  • Supported2
  • NotYetObserved1
By confidence change
  • Unchanged5
  • Decreased1

Learning records needing stronger sourcing: LR_RESTORATION

Learning data gaps (14)
  • — Outcomes vary by enforcement and governance quality.
  • — Outcomes vary by legal recognition and governance context.
  • — Detection alone does not create enforcement.
  • — Recovery is slow and condition-dependent.
  • — Effects depend on pesticide type and farm management.
  • — Area alone is a weak proxy for habitat value.
  • — Quantitative effect sizes vary by region and dataset.
  • — Causal attribution is difficult; context varies.
  • — Response capacity is not modelled here.
  • — No observed time-series in the system yet.
  • — Field vs laboratory effect sizes differ.
  • — Wild vs managed pollinator dynamics not fully mapped.
  • — Quantitative effect sizes vary by region.
  • — Response capacity not modelled.

Learning records are structured examples for the proof cases, not live impact reports.

Evidence coverage
Amazon Living System
Needs review
Claims 4Sources 3

Strong qualitative basis; regional quantification and dataset URLs need verification.

Pollination / Food System
Strong
Claims 7Sources 2

Well-supported overall; varies by crop, region and pollinator group.

Coverage reflects how many claims and sources are attached to each case, and how much remains context-dependent or unverified — not a score of how “good” the underlying science is.

Data quality register (9)
High 0Medium 4Low 5Unverified sources 10Sources needing URL 6
Amazon rainfall and moisture-recycling roles are well-supported qualitatively, but the magnitude varies by region and method.Medium
NeedsQuantificationNode:EC_AMAZON_RAINFOREST
Fix: Add region-specific quantitative ranges with peer-reviewed sources where available.
Pollination dependency is strong overall but varies by crop, geography and pollinator group; it is not uniform.Medium
RegionalVariationNode:FN_POLLINATION
Fix: Differentiate crop- and region-level dependence rather than a single global figure.
Amazon monitoring datasets are referenced at organisation level; specific dataset URLs and access dates still require verification.Low
MissingURLSource:MAPBIOMAS
Fix: Add verified dataset URLs and access dates without fabricating precise report identifiers.
INPE deforestation monitoring is credible at institution level; the precise dataset URL needs verification.Low
MissingURLSource:INPE
Fix: Confirm and add the specific monitoring-programme URL.
A contextual institutional reference without a confirmed URL or precise publication metadata.Medium
IncompleteMetadataSource:AMAZON_INSTITUTIONAL
Fix: Replace with a specific, verifiable report once identified; keep marked until then.
Restoration outcomes depend on method, land-use history and protection from future disturbance; evidence is context-dependent.Medium
ContextDependencySolutionPathway:PW_DEGRAD_RESTORATION
Fix: Keep cautious framing; avoid presenting restoration as an immediate substitute for protection.
A structured learning example derived from existing evidence — not live field data or a 4PLANET implementation record.Low
WeakSourceLearningRecord:LR_RESTORATION
Fix: Connect to live observed data only if and when it exists.
Leverage and urgency are qualitative signals to support reasoning, not a final ranking or automated recommendation.Low
ContextDependencyDecisionSignal:DS_PROTECTED_AREAS
Fix: Preserve qualitative framing; do not present as a definitive priority order.
Internal analysis used for structure and synthesis; not an external authority and should not be cited as primary evidence.Low
WeakSourceSource:INTERNAL
Fix: Back internal synthesis with external sources wherever a claim depends on it.

These are integrity signals — known limitations the system makes visible rather than hides.