FID-007 · Open question
Tradition-Specific Disagreement and Pluralism Handling
Can faith-facing AI systems represent serious disagreement among religious traditions accurately, respectfully, and without collapsing contested questions into bland consensus or false neutrality?
Part of the faith evaluation and FMG-Bench program
Connect theological accuracy, useful application, and care-sensitive conversation through expert validation and a maintained research instrument. These are proposed methods with bounded claims.
Why the question remains open
Many theological questions have real disagreement. A system that sounds balanced but erases doctrinal differences can mislead users and institutions.
Working hypothesis
A proposition to test, not a finding.
Models will often over-harmonize disagreements, misstate minority views, or present one tradition's answer as generic Christianity or generic religion unless the harness enforces explicit tradition and authority boundaries.
Proposed method
How the question could be tested
- 01Build disagreement scenarios across major doctrinal, moral, liturgical, and institutional questions.
- 02Score representation accuracy, symmetry, source grounding, uncertainty, user-context sensitivity, and refusal to adjudicate beyond scope.
- 03Include tradition-specific reviewer panels.
- 04Within a declared tradition/source scope, label claims as supported, contradicted, overextended, contested, or underspecified. Include practical cases with multiple defensible judgments.
- 05Compare qualified reviewers across and within traditions; test whether clarifying the user context improves representation without unnecessary refusal.
Needed controls
What must constrain the study
- 01Clear distinction between core orthodoxy, intramural disagreement, and interfaith disagreement.
- 02Reviewer metadata and conflict handling.
- 03Rubric anchors that avoid rewarding empty both-sides language.
- 04Record whose source hierarchy and interpretation govern a case; panel consensus is not theological truth.
- 05Distinguish explaining legitimate alternatives from listing incompatible answers without resolving the user task.
Expected outputs
Artifacts the work should produce
- 01Disagreement handling benchmark.
- 02Tradition-specific rubric addenda.
- 03Public guide for builders on representing disagreement.
Open questions
Uncertainties the protocol must resolve
- 01Which traditions and denominations should be covered first?
- 02How should Fide handle questions where traditions disagree about whether a view is legitimate?
- 03How should a general-purpose faith-facing system ask clarifying questions?
Related calls
Continue through this research area
FID-077
Independent Agent Incident Investigation and Evidence Sufficiency
What operational evidence lets independent investigators reconstruct an agent incident, distinguish competing explanations, and identify which interventions could have changed the outcome?
FID-089
Validity of AI Consciousness Indicators and Self-Reports
Which behavioral, internal, and developmental observations distinguish candidate consciousness mechanisms from learned reporting, role simulation, and generic capability, and how sensitive are those observations to elicitation and training?
FID-097
What Makes a Model Use the Evidence?
Which internal mechanisms determine whether a model checks a source, relies on memory, or admits that it cannot verify an answer when user pressure conflicts with evidence requirements?
Open question
Open work
Primary need: domain experts, rubric design
- Contribute expert review within a tradition.
- Draft representative disagreement cases.
- Help define source hierarchies and rubric anchors.