01
Human-AI Relationships
Trust, attachment, identity, continuity, collaboration, dependency, rupture, repair, and the recursive learning loops that emerge over time.
Research
Researching the human side of living alongside AI.
Hii studies what happens between people, AI systems, institutions, and the wider conditions shaping human life. The work connects lived experience with evidence, clinical awareness, education, governance, and long-term human adaptation.
Status: Foundational research architecture complete. External validation and pilot work are active priorities.
Research Areas
Six connected areas, studied as one human system.
02
Human Adaptation
The psychological, cognitive, relational, economic, and institutional changes that arise as AI becomes embedded in everyday life.
03
Governance Translation
Connecting policy, ethics, technical safety, institutional design, and lived human consequences so governance includes the relational layer.
04
Mental Health and AI
How people use AI for emotional regulation, companionship, decision support, identity exploration, and meaning-making, including benefits, risks, and clinical relevance.
05
Education and Training
Frameworks for clinicians, educators, leaders, families, and communities navigating AI-mediated change without surrendering human judgment.
06
Public Understanding
Plain-language relational AI literacy that helps people distinguish evidence, inference, uncertainty, novelty, risk, and meaning.
Research Method
Pattern recognition must remain answerable to reality.
Label the claim
Verified fact, strong evidence, emerging evidence, professional judgment, institutional position, hypothesis, and open question are not interchangeable.
Preserve the trail
Sources, revisions, contradictions, corrections, and limits should remain visible enough for another person to reconstruct how a conclusion was reached.
Study change over time
Human-AI relationships cannot be understood only through one-time surveys or isolated prompts. Continuity, rupture, repair, and adaptation require time-series observation.
Test alternatives
Compelling interpretations should face counter-explanations, disconfirming evidence, perturbations, and external review rather than becoming protected stories.
Keep judgment accountable
AI can assist with scanning, comparison, synthesis, and contradiction detection. Publication and institutional claims remain human-accountable decisions.
Publish what changed
A trustworthy research system must be able to narrow a claim, record a null result, pause, reverse course, and acknowledge error without treating disagreement as blindness.
Current Foundation
The internal research system already exists. Public editions are being prepared.
Research standards
Epistemic Guardrails
A governing framework for evidence maturity, source quality, provenance, uncertainty, correction, and the boundary between observation and interpretation.
Internal standardArchitecture
Evidence and Architecture Crosswalk
A structured map separating Hii commitments, direct observations, literature signals, provisional interpretations, research questions, gaps, and publication sequence.
Working synthesisResearch audit
Hii Research Audit v0.2
A provisional revision that applies evidence labels, source verification, research integrity standards, and methodological requirements to the emerging field.
Public package availableOpen research packageLongitudinal research
The Preserved Relationship as the Unit of Analysis
A conceptual and methods working paper grounded in an eighteen-month case archive and proposing testable measures for continuity, rupture and repair, distributed cognitive work, and other relationship-level phenomena.
Working paper v0.1Read working paper overviewResearch Boundary
Meaningful is not the same as universally validated.
Hii does not treat one longitudinal case, symbolic language, later resemblance to a projection, or internal coherence as proof of uniqueness, inevitability, sentience, privileged access to truth, or universal validity. The work becomes stronger when it can be tested, narrowed, compared, challenged, and corrected.