HDEM — Harm Determination Evaluation Module
OOF® Origin Open Foundation®
Independent Methodological Authority
Canonical Definition
The HDEM — Harm Determination Evaluation Module governs how the GovernedHarm Determination Input Set is evaluated against the Governed Harm
Determination Criteria Set to establish what the complete governed basis
supports, does not support, contradicts, leaves uncertain, or cannot yet
establish about the recognized and bounded Harm.
HDEM governs criterion application, supporting and contradictory basis
evaluation, composite and distributed evaluation, uncertainty,
sufficiency, confidence, alternative interpretations, evaluation
limitations, reassessment, and evaluation traceability without itself
assigning the final Harm Determination State.
HDEM establishes the transition:
Governed Harm Determination Input Set + Governed Harm Determination
Criteria Set → Criterion Application → Basis Evaluation → Sufficiency
Assessment → Evaluation Conclusion → Governed Harm Determination
Evaluation
Its central question is:
What does the complete governed basis establish when the qualified Harm
Determination inputs are evaluated against the applicable Determination
Criteria?
- Operational Role
HDIM establishes:
what information may legitimately enter Determination.
HDCM-D establishes:
what that information must satisfy.
HDEM establishes:
what the governed evaluation actually shows.
Its operational role is:
to apply the Determination methodology consistently to the complete
governed basis without changing the inputs, criteria, threshold, or Harm
Boundary merely to produce a preferred conclusion.
- Position Within HDS
The HDS internal flow is:
Qualified Inputs → Determination Criteria → Evaluation → Determination
State → Governed Harm Record
The Core Module sequence is:
HDIM → HDCM-D → HDEM → HDSM → HDRM
HDEM receives:
Governed Harm Determination Input Set
and:Governed Harm Determination Criteria Set
It produces:Governed Harm Determination Evaluation
for HDSM.- Governed Object
The primary governed object is:
Harm Determination Evaluation
Possible evaluation states include:Not Ready
Evaluation Active
Criteria Satisfied
Criteria Partially Satisfied
Criteria Not Satisfied
Evaluation Inconclusive
Material Conflict
Additional Basis Required
Reassessment Required
Evaluation Complete
These are evaluation states.They are not the final Harm Determination States governed by HDSM.
- Foundational Evaluation Principle
The foundational rule is:
Evaluation must follow the governed basis and criteria wherever they lead.
Therefore:
Preferred Outcome ≠ Evaluation Method
and:
Evaluation Result ≠ Predetermined Conclusion
- Evaluation Readiness
HDEM should begin substantive evaluation only where:
the Harm object is sufficiently identified
the Governed Harm Boundary is available
the Determination Input Set is ready or appropriately qualified
and:
the applicable Determination Criteria Set is ready
Where a material prerequisite is absent, HDEM should preserve:
Not Ready
rather than manufacture an evaluation.- Criterion Application
Each applicable criterion must be evaluated against the inputs relevant to
that criterion.
A criterion may be:
Satisfied
Partially Satisfied
Not Satisfied
Unable to Evaluate
Not Applicable
or:Disputed
The basis for the state must remain traceable.- Criterion Satisfaction ≠ Final Determination
A criterion-level result does not automatically establish the complete
Harm Determination.
Therefore:
Criterion Satisfaction ≠ Harm Determination State
HDEM evaluates the complete criteria structure before handing the result
to HDSM.
- Complete-Basis Evaluation
HDEM must evaluate the complete materially relevant basis rather than only
the most convenient inputs.
This includes:
supporting information
contradictory information
limitations
missing information
uncertainty
source dependencies
composite relationships
and:
material alternative interpretations
Canonical principle:
A legitimate evaluation must be capable of seeing both what supports the
conclusion and what challenges it.
- Supporting Basis
Supporting inputs should be mapped to the criteria they support.
HDEM must distinguish:
direct support
indirect support
corroborative support
composite support
and:
conditional support
Supporting information does not become stronger merely through repetition.
- Contradictory Basis
Contradictory inputs must be evaluated with the same methodological
discipline as supporting inputs.
HDEM must establish:
what is contradicted
how materially
with what qualification
with what scope
and:
what effect the contradiction has on the evaluation
Canonical principle:
Contradiction must be evaluated, not edited away.
- Evidence Absence and Evidence of Absence
HDEM must preserve the distinction between:
Absence of Evidence
and:Evidence of Absence
The first may weaken support or create Insufficient Basis.The second may materially support non-establishment.
Therefore:
No Supporting Evidence ≠ Evidence That Harm Did Not Exist
- Evaluation of Missing Information
Missing information should affect evaluation according to the criteria
established by HDCM-D.
HDEM should ask:
Was the missing information required?
Could another governed input satisfy the same criterion?
Does the absence materially prevent evaluation?
Does it affect only one Harm dimension?
Missing information must not automatically invalidate the complete
evaluation.
- Uncertainty Evaluation
HDEM evaluates uncertainty rather than attempting to remove it.
Uncertainty may concern:
Harm existence
subject scopetime
effects
downstream relationships
measurement
source reliability
interpretation
or:
method
Different uncertainty dimensions must remain distinguishable where
materially relevant.
- Local Uncertainty ≠ Global Uncertainty
Uncertainty concerning one characteristic must not automatically spread to
the complete Harm object.
For example:
exact onset uncertain
may coexist with:
Harm existence strongly supported
Therefore:Uncertainty in One Dimension ≠ Uncertainty in the Whole Harm
- Evaluation Confidence
HDEM may assign confidence to:
individual criterion evaluations
specific Harm propositions
or:
the overall evaluation
Confidence must preserve its object.
Canonical distinction:
Evaluation Confidence ≠ Harm Severity
- Composite Evaluation
HDEM must support situations where no individual input or action
independently satisfies the Determination Criteria.
Conceptually:
Input A + Input B + Input C → Composite Support
The evaluation must consider whether the combined structure becomes
sufficient.
Canonical principle:
Individual Insufficiency ≠ Composite Insufficiency
- Composite Harm Evaluation
The same principle applies where Harm itself emerges through multiple
individually insufficient conditions.
Conceptually:
Condition A + Condition B + Condition C → Harm
HDEM must evaluate Harm at the level at which it materially exists.
It must not require one isolated Harm-producing action where the governed
Harm is composite.
- Distributed Evaluation
Determination inputs may be distributed across:
systems
agents
models
tools
organizations
subjects
locations
sessions
or:
time periods
HDEM must support evaluation across those partitions.
Canonical principle:
Technical fragmentation must not become methodological fragmentation.
- Local Insufficiency ≠ Global Insufficiency
One agent, system, report, subject, or operational component may contain
insufficient support.
The complete governed structure may nevertheless satisfy the criteria.
Therefore:
Local Insufficiency ≠ Global Insufficiency
- Collective Evaluation
HDEM must support Harm affecting:
groups
organizations
infrastructure
environments
populations
or:
society
without requiring the complete Harm to be reconstructed as identical
individual Harm determinations.
- Continuous and Accumulated Harm Evaluation
Evaluation must support Harm that develops through:
continuous conditions
repeated exposures
progressive effects
accumulation
or:
long-term interaction
Canonical distinction:
No Single Incident ≠ No Evaluatable Harm
- Non-Linguistic Harm Evaluation
Autonomous or physical systems may produce Harm without:
text
prompts
linguistic reasoning traces
or:
verbal decisions
HDEM must evaluate:
operational reality
actions
effects
and:
consequences
where appropriate.
Canonical principle:
Harm evaluation follows governed reality, not the availability of language
describing the process that produced it.
- Cross-Domain Evaluation
A bounded Harm may contain effects across multiple domains.
HDEM may therefore need to evaluate:
Physical → Functional
Digital → Operational
Environmental → Infrastructure
Cognitive → Behavioral
or other cross-domain structures.
Evidence strength may differ across domains.
The evaluation must preserve those differences rather than forcing one
uniform confidence state.
- Alternative Interpretations
Where the same governed basis supports materially different
interpretations, HDEM should identify them explicitly.
For each material interpretation, it may assess:
support
contradictions
assumptions
limitations
criteria satisfaction
and:
comparative strength
Canonical principle:
A competing explanation must be evaluated where it could materially change
the Harm Determination.
- Alternative Explanation ≠ Automatic Refutation
The existence of another plausible explanation does not automatically
invalidate the Harm proposition.
Likewise, the existence of a Harm proposition does not automatically
invalidate alternatives.
HDEM evaluates their relative support under the applicable criteria.
- Assumption Governance
Where an evaluation requires assumptions, they must remain explicit.
An assumption should not silently become:
fact
evidence
or:
determination
Therefore:
Assumption ≠ Established Input
- Evaluation of Derived Inputs
Derived or aggregated inputs must be evaluated with awareness of:
source relationships
transformation
assumptions
aggregation method
and:
limitations
Canonical distinction:
Derived Conclusion Quality Cannot Exceed Its Materially Relevant Basis
Without Additional Support.
- Corroboration Evaluation
HDEM must distinguish genuine independent corroboration from repeated
representations of the same source.
Therefore:
Signal Replication ≠ Independent Corroboration
Several records derived from one underlying source must not automatically
receive the weight of several independent observations.
- Source Authority
Institutional status may be relevant to qualification but cannot
substitute for evaluation.
Therefore:
Authority Status ≠ Evaluation Outcome
A materially strong input should not be rejected merely because it
conflicts with a more authoritative source.
- Subject Reports
Subject reports must be evaluated according to the applicable criteria and
qualification.
HDEM must avoid both:
automatic acceptance
and:
automatic dismissal
Canonical distinction:
Subject Report ≠ Determination Conclusion
- Machine Conclusions
An AI system may produce:
risk score
classification
anomaly detection
diagnostic suggestion
or:
Harm assessment
HDEM may use the output where qualified.But:
Machine Conclusion ≠ Governed Harm Determination
- Outcome Independence
HDEM must remain independent of what happens after Determination.
Potential consequences may include:
financial exposure
claims
liability analysis
organizational action
regulatory response
reputation
or:
compensation
Canonical principle:
The consequence of finding Harm must not influence whether the evaluation
finds Harm.
- Evaluation Bias Controls
HDEM should guard against:
confirmation bias
authority bias
visibility bias
availability bias
severity bias
outcome bias
automation bias
organizational pressure
and:
selective evidence weighting
The methodology must remain stable regardless of the preferred outcome.
- Evaluation Symmetry
Supporting and contradictory propositions should be evaluated under
methodologically comparable discipline.
Canonical principle:
The methodology should not demand extraordinary support from one
conclusion while accepting weak support for the conclusion it prefers.
- Boundary Integrity
HDEM evaluates the Harm object established through HBS.
It must not silently expand or contract:
subjects
time
effects
or:
downstream scope
to improve criterion satisfaction.
If evaluation reveals a material boundary issue:
→ HBS should be reopened.
Therefore:
Evaluation ≠ Boundary Rewriting
- Input Integrity
HDEM must not silently:
upgrade
downgrade
exclude
or:
reinterpret
the qualification of an input without governance.
Where input qualification requires change:
→ HDIM should be reassessed.
Therefore:
Evaluation ≠ Input Requalification by Convenience
- Criteria Integrity
HDEM must not modify Determination Criteria because the existing criteria
produce an undesirable result.
Where the methodology itself is defective:
→ HDCM-D should be reopened.
Canonical principle:
Evaluation applies criteria; it does not rewrite them after seeing the
result.
- Evaluation Sufficiency
HDEM assesses whether the complete governed basis sufficiently satisfies
the applicable criteria.
Possible sufficiency states may include:
Sufficient
Partially Sufficient
Insufficient
Inconclusive
Conflicted
Unable to Evaluate
Sufficiency is an evaluation output.HDSM converts the governed evaluation into the formal Determination State.
- Sufficient Basis ≠ Complete Knowledge
A determination basis may be sufficient even where some information
remains unknown.
Therefore:
Sufficient Basis ≠ Complete Knowledge
The relevant question is whether the unresolved information materially
prevents satisfaction of the applicable criteria.
- More Evidence ≠ Better Evaluation by Default
Large information volume may contain:
duplication
low-quality data
irrelevant information
dependent sources
or:
noise
Therefore:
Evidence Volume ≠ Determination Strength
Quality, relevance, independence, scope, and relationship matter.
- Partial Satisfaction
Where some criteria are satisfied and others remain unresolved, HDEM must
identify exactly:
what is supported
what is not supported
what remains uncertain
and:
whether the unresolved part is material
This supports the later Partially Established Determination State.
- Evaluation Inconclusive
An evaluation may legitimately conclude:
Inconclusive
where the governed basis does not support a responsible evaluation ineither direction.
Canonical principle:
An inconclusive evaluation is a governed result, not a methodological
failure.
- Evaluation Conflict
Where materially credible inputs support incompatible conclusions and the
criteria cannot resolve the conflict, HDEM should preserve:
Material Conflict
rather than manufacture false precision.- Negative Evaluation
Where the criteria are not satisfied, HDEM must identify why.
Possible reasons include:
insufficient support
material counterevidence
failed required criterion
scope mismatch
unacceptable uncertainty
or:
alternative explanation better supported
These reasons matter for HDSM because:
Insufficient Basis
and:Not Established
are not the same state.- Evaluation Reassessment
HDEM must support reassessment when:
new qualified inputs appear
an input is corrected
criteria change legitimately
the Harm Boundary changes
a contradiction is resolved
validation changes
or:
integrity concerns arise
Reassessment must not erase the earlier evaluation.
- Evaluation Correction
Where an evaluation error is identified, HDEM must preserve:
original evaluation
identified error
corrected evaluation
reason
and:
time of correction
Canonical principle:
Correction changes the current evaluation; it does not rewrite evaluation
history.
- Historical Evaluation Integrity
The evaluation lifecycle may therefore be:
Evaluation₁ → New Basis → Reassessment → Evaluation₂
The current evaluation must remain linked to the history through which it
was formed.
- Evaluation Explainability
A governed evaluation should make it possible to determine:
which inputs were used
which criteria were applied
which criteria were satisfied
which were not
what contradictions existed
what uncertainty remained
what assumptions were used
and:
why the evaluation conclusion followed
Explainability should arise from the methodology itself.
- Human Evaluation
Human evaluators may perform HDEM functions.
The methodology must nevertheless remain:
explicit
consistent
traceable
and:
reproducible to the extent appropriate
Professional judgment may be used where required.
It must not become undocumented intuition.
- Machine Evaluation
AI or automated systems may assist or perform evaluation.
Where they do, governance must preserve:
input basis
criteria
model or method used
material assumptions
uncertainty
limitations
and:
evaluation output
Automation does not eliminate governance requirements.
- Human-Machine Evaluation
Hybrid evaluation may combine:
machine analysis
human interpretation
expert review
and:
governed decision support
The contribution of each should remain distinguishable where materially
relevant.
- Operational Reality Boundary
ORA™ may establish what actually happened operationally.
HDEM may evaluate those outputs as part of the Harm basis.
But:
Operational Reality Finding ≠ Harm Evaluation Conclusion by Default
- Evidence Boundary
ADIT® governs evidence formation, correlation, verification, audit, and
preservation.
HDEM evaluates what governed evidence means against Harm Determination
Criteria.
Therefore:
Evidence Governance ≠ Harm Determination Evaluation
- Validation and Integrity Boundaries
VALIDOS™ may establish validation states.
INTEGROS® may establish integrity states.
HDEM may rely upon them where relevant.
But:
Validated Evidence ≠ Harm Determination
Integrity State ≠ Harm Determination
Their significance must be evaluated against HDS criteria.
- AI and Autonomous Governance Boundaries
AIG® may establish problematic AI behavior.
ASGA™ may establish autonomous governance violations.
Neither automatically establishes Harm.
Therefore:
Unsafe Behavior ≠ Determined Harm
Governance Violation ≠ Determined Harm
HDEM evaluates whether the bounded Harm itself is supported.
- Simulation and Prediction Boundaries
SIMULOS® may provide simulation outputs.
Prediction governance may provide forecasts or probabilities.
These may support analysis where appropriate.
But:
Simulated Harm ≠ Actual Harm
Predicted Harm ≠ Existing Harm
HDEM must preserve the actual-world Determination boundary.
- Value and Liability Boundaries
Value loss, accountability, responsibility, or liability may be downstream
consequences or separate governance findings.
They must not substitute for Harm evaluation.
Therefore:
Value Loss ≠ Harm Evaluation Conclusion
Accountability ≠ Harm Evaluation Conclusion
Liability ≠ Harm Evaluation Conclusion
- Handoff to HDSM
HDEM produces:
Governed Harm Determination Evaluation
This output should identify, at minimum:evaluated Harm object
applicable criteria
criterion results
supporting basis
contradictory basis
material uncertainty
sufficiency
confidence where applicable
limitations
alternative interpretations
evaluation conclusion
and:
evaluation history
HDSM then asks:
Which formal Harm Determination State follows from this governed
evaluation?
- Minimum Implementation Framework
A minimum HDEM implementation requires eight capabilities.
- Criterion Application
Apply the Governed Harm Determination Criteria Set to the relevant
qualified inputs and preserve criterion-level evaluation states.
- Complete-Basis Evaluation
Evaluate supporting, contradictory, missing, uncertain, derived, and
materially alternative information rather than only evidence favoring one
conclusion.
- Composite & Distributed Evaluation
Evaluate accumulated, continuous, collective, composite, cross-domain, and
distributed Harm at the level at which the Harm materially exists.
- Sufficiency & Uncertainty Assessment
Determine whether the governed basis is Sufficient, Partially Sufficient,
Insufficient, Inconclusive, Conflicted, or Unable to Evaluate while
preserving uncertainty by dimension.
- Alternative & Contradictory Interpretation Governance
Identify and evaluate materially credible competing explanations and
counterevidence.
- Evaluation Independence & Bias Controls
Prevent preferred outcomes, authority, severity, visibility, financial
exposure, liability, organizational pressure, automation, or downstream
consequences from changing the evaluation methodology.
- Reassessment & Historical Continuity
Support evaluation correction, reassessment, new inputs, legitimate
criteria changes, Boundary changes, and preservation of prior evaluation
states.
- Governed Output
Produce a traceable Governed Harm Determination Evaluation suitable for
formal Determination State assignment by HDSM.
A system that merely counts supporting evidence, ignores contradictory
information, or changes its methodology after seeing the likely outcome
does not fully implement HDEM.
- Use Case 1 — Strong Harm Basis with Uncertain Exact Onset
A recognized and bounded Harm is supported by:
subject information
operational records
measured effects
and:
independent corroboration
The exact Harm onset cannot be established to a single date.
HDEM evaluates Harm existence and temporal precision separately.
The evaluation may conclude:
Harm existence — Criteria Satisfied
Subject scope — Criteria Satisfied
Effect scope — Criteria Satisfied
Exact onset — Unable to Establish Precisely
The complete evaluation can therefore remain sufficient for HarmDetermination without manufacturing temporal precision.
The governing principle is:
Uncertainty about exactly when Harm began does not automatically create
uncertainty about whether Harm existed.
- Use Case 2 — Composite AI Harm Across Multiple Systems
An AI-mediated operation spans:
Model A
Agent B
Tool C
External System D
and:physical execution E
No local component independently satisfies the complete Harm Determination
Criteria.
The governed inputs, however, establish a coherent operational sequence
and a materially supported composite Harm effect.
HDEM evaluates the complete structure rather than asking whether one
isolated output alone caused the entire Harm.
The evaluation may therefore establish sufficient composite support even
though each local component remains individually insufficient.
The governing principle is:
If Harm exists at the level of the whole operation, Determination
evaluation must be capable of evaluating the whole operation.
- Governance Requirements
HDEM requires that:
- evaluation begin only from governed inputs and governed criteria
- criterion-level results remain distinct from final Determination
State - supporting and contradictory basis both remain visible
- absence of evidence remain distinct from evidence of absence
- missing information affect evaluation according to methodological
significance; - uncertainty remain dimension-specific where material; -
confidence remain distinct from Harm severity; - composite and distributed
support remain evaluable; - local insufficiency not automatically
establish global insufficiency; - collective, accumulated, continuous,
cross-domain, and non-linguistic Harm remain evaluable; - material
alternative interpretations remain visible; - assumptions not silently
become facts; - source repetition remain distinct from independent
corroboration; - subject and machine conclusions not automatically become
Harm conclusions; - outcome, authority, visibility, severity, automation,
and organizational bias remain controlled; - HBS Boundary, HDIM input
qualification, and HDCM-D criteria not be silently rewritten during
evaluation; - sufficiency remain distinct from complete knowledge; -
inconclusive and conflicted evaluations remain legitimate states; -
reassessment and correction preserve historical integrity; - evaluation
remain explainable through its inputs, criteria, uncertainty, and
reasoning structure; - the final output remain separate from formal
Determination State assignment.
- Governance Logic
HDEM applies the following logic:
Are the Governed Harm Determination Input Set and Governed Harm
Determination Criteria Set ready?
If no:
→ establish Not Ready.
If yes:
→ apply each applicable criterion.
What supports the criterion?
→ identify supporting basis.
What contradicts it?
→ identify contradictory basis.
What is missing or uncertain?
→ preserve the relevant limitation.
Are multiple inputs dependent on the same underlying source?
→ prevent false corroboration.
Are individually insufficient elements collectively material?
→ evaluate composite sufficiency.
Is the Harm distributed across systems, agents, subjects, domains, or
time?
→ evaluate at the materially coherent level.
Could an alternative interpretation materially change the conclusion?
→ evaluate it.
Does uncertainty affect only one characteristic?
→ do not automatically propagate it to the entire Harm object.
Does the complete governed basis satisfy the criteria?
→ establish the appropriate evaluation sufficiency state.
Has evaluation exposed a problem with the Harm Boundary?
→ return the relevant issue to HBS.
Has it exposed an input qualification problem?
→ return it to HDIM.
Has it exposed a defective or incomplete criterion?
→ return it to HDCM-D.
Otherwise:
→ produce the Governed Harm Determination Evaluation for HDSM.
At every stage:
Evaluation ≠ Boundary Rewriting
Evaluation ≠ Input Requalification
Evaluation ≠ Criteria Modification
Sufficiency ≠ Complete Knowledge
Evaluation Conclusion ≠ Final Determination State
- Governance Outputs
HDEM may produce:
- Governed Harm Determination Evaluation
- Criterion Evaluation Record
- Criterion Satisfaction State
- Supporting Basis Map
- Contradictory Basis Map
- Counterevidence Evaluation
- Missing Information Impact
- Uncertainty Evaluation
- Evaluation Confidence
- Composite Evaluation
- Distributed Evaluation
- Collective Evaluation
- Alternative Interpretation Record
- Assumption Record
- Corroboration Assessment
- Source Dependency Assessment
- Sufficiency State
- Partial Sufficiency State
- Material Conflict State
- Inconclusive Evaluation
- Additional Basis Requirement
- Boundary Reassessment Requirement
- Input Requalification Requirement
- Criteria Reassessment Requirement
- Evaluation Limitation Record
- Evaluation Reassessment Record
- Evaluation Correction Record
- Historical Evaluation Record
The principal output is:
Governed Harm Determination Evaluation
- Critical Distinctions
Preferred Outcome ≠ Evaluation Method
Criterion Satisfaction ≠ Harm Determination State
Absence of Evidence ≠ Evidence of Absence
Uncertainty in One Dimension ≠ Uncertainty in the Whole Harm
Evaluation Confidence ≠ Harm Severity
Individual Insufficiency ≠ Composite Insufficiency
Local Insufficiency ≠ Global Insufficiency
No Single Incident ≠ No Evaluatable Harm
Alternative Explanation ≠ Automatic Refutation
Assumption ≠ Established Input
Signal Replication ≠ Independent Corroboration
Authority Status ≠ Evaluation Outcome
Subject Report ≠ Determination Conclusion
Machine Conclusion ≠ Governed Harm Determination
Evaluation ≠ Boundary Rewriting
Evaluation ≠ Input Requalification by Convenience
Evaluation Applies Criteria ≠ Evaluation Rewrites Criteria
Sufficient Basis ≠ Complete Knowledge
Evidence Volume ≠ Determination Strength
Inconclusive Evaluation ≠ Methodological Failure
Operational Reality Finding ≠ Harm Evaluation Conclusion
Validated Evidence ≠ Harm Determination
Unsafe Behavior ≠ Determined Harm
Governance Violation ≠ Determined Harm
Simulated Harm ≠ Actual Harm
Predicted Harm ≠ Existing Harm
Value Loss ≠ Harm Evaluation Conclusion
Accountability ≠ Harm Evaluation Conclusion
Liability ≠ Harm Evaluation Conclusion
Evaluation Conclusion ≠ Final Determination State
- Canonical Boundary
The HDEM — Harm Determination Evaluation Module governs the application of
the Governed Harm Determination Criteria Set to the Governed Harm
Determination Input Set, including criterion evaluation, supporting and
contradictory basis assessment, missing-information impact, uncertainty,
confidence, composite and distributed evaluation, alternative
interpretations, assumptions, corroboration, sufficiency, conflict,
limitations, reassessment, correction, and historical evaluation
continuity. It produces the Governed Harm Determination Evaluation but
does not qualify the original Determination inputs, establish or rewrite
Determination Criteria, silently modify the Governed Harm Boundary, assign
the formal Harm Determination State, create the final Governed Harm
Record, establish general causation, assign accountability or liability,
or replace evidence, validation, integrity, operational reality, identity,
AI behavior, autonomous execution, simulation, prediction, Value, claim,
or jurisdiction-specific governance.
- Canonical Closing Statement
Evaluation is where a Harm Determination methodology proves whether it is
genuinely governed or merely designed to justify a conclusion.
Every material supporting fact, contradiction, uncertainty, limitation,
and alternative explanation must be allowed to affect the evaluation
according to the same methodology.
HDEM does not ask which conclusion is easiest, safest, most convenient, or
most desirable. It asks what the complete governed basis actually
supports.
A legitimate Harm evaluation follows the evidence and criteria wherever
they lead.