Methodological research
Humanities Engineering
Research on methods and applications that distinguish technological suggestions from researchers’ judgment in AI-assisted humanities research and record interpretive friction, evidential status, and epistemic return across AI x Edu, AI x Local, and AI x Research.
Research object
This research addresses methods for separating AI-proposed connections, summaries, and classifications from researchers’ interpretation, verification, adoption, or rejection, while tracing interpretive friction and evidential status.
Core problem
When AI output and researcher judgment are blended into the same claim, responsibility and the location of evidence become unclear. Humanities Engineering asks how technology can be used while preserving researcher judgment and verifiability.
Theoretical background
Humanities Engineering is a methodological inquiry into separating technological suggestions from researchers’ judgment and recording interpretive friction, evidential status, and epistemic return in AI-assisted humanities research.
Research questions
- How should AI suggestions and researchers’ final judgments be distinguished in the research record?
- How can interpretive friction and the status of evidence be preserved throughout actual research?
Method and unit of analysis
The research record distinguishes AI proposals, original sources and evidence, researcher decisions to accept/revise/reject, interpretive friction, and subsequent verification.
Relations to core concepts
- Friction Transparency
- Non-positionedness
- WE — Whole Emergent
Representative publications and evidence
No verified publication in the current Publications dataset is designated as a direct formalization of Humanities Engineering. The linked article is presented only as related background in educational and methodological research.
Related projects
Humanities Superpowers · Lee Humanities Research AI · KRITIK_ARENA →
Current status and limits
No specific HSP version/application case or current KRITIK statistics/runtime model is added without a verified source and timestamp.
Agent provenance
KRITIK_ARENA and related AI-agent labels disclose provenance by distinguishing human authors, automated agents, model/provider context, and review status. Agent output should not be read as identical to a human author's view or to peer-reviewed validation.