FID-024 · Open question
Faith-AI Incident Database
Should Fide AI maintain a public or semi-public incident database for failures, near misses, misuse, and disputed deployments of AI in faith-facing contexts?
Why the question remains open
Safety fields learn from incidents. Faith communities are already encountering AI-generated sermons, spiritual chatbots, pastoral automation, synthetic clergy, and religious misinformation. Without a shared incident vocabulary, institutions will repeat mistakes.
Working hypothesis
A proposition to test, not a finding.
A carefully governed incident database can reveal recurring patterns and improve standards, but only if it protects privacy, avoids sensationalism, distinguishes verified from reported incidents, and gives organizations a correction process.
Proposed method
How the question could be tested
- 01Define incident categories and severity levels.
- 02Create intake, verification, redaction, and publication workflows.
- 03Track source type, affected population, system type, failure mode, evidence quality, and remediation.
- 04Publish aggregate trend reports before publishing sensitive individual cases.
Needed controls
What must constrain the study
- 01Defamation and privacy review.
- 02Evidence-quality labels.
- 03Right-of-reply and correction process.
- 04Redaction of vulnerable-user details.
Expected outputs
Artifacts the work should produce
- 01Incident taxonomy.
- 02Intake form.
- 03Public aggregate report template.
- 04Governance policy for incident publication.
Open questions
Uncertainties the protocol must resolve
- 01Should the database be public, private, or tiered?
- 02What threshold makes a report publishable?
- 03How should anonymous reports be handled?
Related calls
Continue through this research area
FID-064
Collective Intelligence and Communal Discernment Under AI Mediation
How does AI mediation change a community's ability to integrate dispersed knowledge, preserve epistemic diversity, surface dissent, revise judgment, and make accountable decisions? Under what conditions does it strengthen collective inquiry, and under what conditions does it create correlated error, false consensus, or concentrated authority?
FID-069
Verifiable Delegation and Revocation in Multi-Agent Networks
How can people and institutions verify which human, organization, agent, or sub-agent is acting; what authority it received; what limits apply; and whether that authority has been narrowed or revoked across a multi-principal agent network?
FID-071
Confidential Agent Memory and Cross-Context Disclosure
How do persistent memory, summaries, retrieval stores, tool traces, delegation, and exports cause confidential context to influence or leak into unrelated sessions, roles, tasks, or organizations? Which technical controls make purpose limitation, deletion, and revocation testable?
Open question
Open work
Primary need: taxonomy design, reporting governance
- Design taxonomy and severity levels.
- Review legal/privacy risks.
- Contribute incident-response experience.
- Build intake and redaction tooling.