Evidence unit
Each central claim is connected to a date, person, place, and primary or reliable secondary source. Firsts, numbers, and causal claims receive extra scrutiny.
Humility before the record
The museum does not use history as a sermon illustration. It distinguishes what the record says, what the museum interprets, and what remains unknown.
Each central claim is connected to a date, person, place, and primary or reliable secondary source. Firsts, numbers, and causal claims receive extra scrutiny.
Happened afterward does not mean happened because of. Sequence, influence, collaboration, and direct causation are treated differently.
Missionary records do not stand in for Korean experience. Wherever possible, Korean authors, women, laypeople, and local communities are sought alongside them.
Church action is not automatically labeled biblical impact. Biblical text, interpretation, institutional behavior, and cultural habit are distinguished.
Visitors may report factual errors, translation problems, missing attribution, or misleading language. Submissions are reviewed, and material corrections will record what changed and why. Sensitive reports involving living people or communities will be handled with care before publication.
billmajors@crossculture-community.comThe Search for Identity · AI-guided oral histories
The planned system will let a participant begin in Korean or English while AI conducts the interview, asks relevant follow-up questions, prepares drafts in both languages, and requests confirmation of names and dates. Final authority belongs to the person whose story is told—not to the museum founder.
Choose Korean or English and separately allow or decline recording, portrait, identification, translation, and research use.
AI follows a governed question framework, asks follow-ups based on the answers, and must not invent missing facts.
See the source transcript beside Korean and English drafts, with translation uncertainty and names or dates needing verification clearly marked.
The participant corrects wording, answers again when desired, and removes anything that should not be published.
Only safety concerns, allegations about others, historical conflicts, legal or rights issues, and serious translation uncertainty go to human review.
Only participant-approved exhibit cards may be released; the participant chooses publication timing and how their name appears.
This page states the process the museum intends to provide. AI interviews will not open until secure storage, privacy controls, consent records, and participant approval functions are working.
Interview development will resume after the core museum is complete.Security before collection
Account verification protects participants and limits repeat interviews. An email address or legal name does not become the story’s identifier. The system uses an internal participant number, and access decisions happen on the server.
Private testimony, working translations, restricted review notes, and the final public exhibit remain separate records.
Recording, translation, identification, portrait, research, and publication are separate choices. Changes create a new record without erasing what was previously chosen.
An unfinished interview or an approved translation is not publication consent. Only the exact artifact approved by the participant can be released.
Participants must be able to download their records, withdraw publication, and request deletion without contacting Bill Majors personally.
Current status: a limited text-interview pilot with secure storage and consent records is being prepared. AI follow-up questions, participant accounts, and public recruitment remain off.
External references are labeled English or 한국어. Korean primary sources remain visible, while English counterparts are added where available.