Continuity
Recognition across change
How archives, memory systems, language, expectations, and repair practices create or fail to create continuity across model and product changes.
Research Area
Human-AI Relationships
Hii studies what changes when interaction with AI becomes sustained, personal, recursive, and consequential rather than occasional tool use.
Focus: trust, attachment, identity, continuity, dependency, co-adaptation, rupture, repair, and distributed cognitive work.
Central Question
What becomes visible when the relationship, not only the prompt, is studied over time?
Single interactions can miss accumulated trust, changing use, model migration, continuity failure, repair, spillover into human relationships, and the ways people and systems adapt to one another.
Research Themes
The relationship is a changing system.
Agency
Support or erosion
When sustained AI use expands reflection, learning, and action, and when it narrows judgment, increases dependency, or displaces human connection.
Co-adaptation
Recursive learning
How humans change their thinking and behavior in response to AI, while AI responses change through context, memory, feedback, and system updates.
Rupture and repair
Sequence matters
How discontinuity, contradiction, memory failure, tone change, and re-alignment affect trust and function across time.
Current Work
From lived archive to testable research program.
Working paper
The Preserved Relationship as the Unit of Analysis
A conceptual and methods paper proposing relationship-level measures, comparison conditions, perturbations, and independent coding.
Read the overviewResearch standard
Epistemic Guardrails
The method requires ordinary explanations first, visible uncertainty, independent review, and publishable null results.
View the standardBoundary
Relationship-level research does not require a claim about machine consciousness.
The research object is the interaction system across the human participant, changing AI instances, archives, interfaces, social context, and time. Claims remain open to alternative explanations and external testing.