Human-AI Research Continuity Theory (HARCT)
Long-term Human-AI research collaboration is fundamentally a problem of research continuity rather than intelligence alone.
GitHub Pages: 논문 온라인 보기
Long-term Human-AI collaboration presents challenges that extend beyond the capabilities of either human memory or AI memory alone. This study investigates long-term Human-AI research collaboration through a multi-repository research program and proposes Human-AI Research Continuity Theory (HARCT) as a unified explanatory framework: long-term Human-AI research collaboration is fundamentally a problem of research continuity rather than intelligence alone.
HARCT is supported by four theoretical components. Goal Preservation Theory argues that preserving research goals is more fundamental than preserving information — goals enable reconstruction, while information without direction becomes an unusable archive. Cognitive Architecture Theory identifies a three-component distributed system in which humans preserve goals, AI systems perform reasoning and reconstruction, and artifacts preserve context. Reconstruction Cost Theory frames documentation quality as an economic variable that directly determines collaboration efficiency. Externalized Memory describes how context is stored outside active participants through persistent artifacts.
Human-AI Collaboration, Human-AI Research Continuity Theory (HARCT), Goal Preservation, Reconstruction Cost, Cognitive Labor Division, Externalized Memory, Distributed Research Memory, Context Transfer, Long-Term Research Programs, Future-Self Collaboration, AI-to-AI Handover
This repository is Paper 5 — Methodological Meta-layer in the Structure Recognition Research Program.
Paper 1: KMap Structure Invariance (empirical case 1 — visual pattern discovery)
Paper 2: Symmetric Boolean Functions (empirical case 2 — structural regularity)
Paper 3: Variable Rearrangement (empirical case 3 — structure under transformation)
↓
Paper 4: Structure Recognition Theory (theoretical hub — unifying Papers 1–3)
↓
Paper 5: Human-AI Research Collaboration ← This repository
(methodological meta-layer — studying the process that produced Papers 1–4)
Paper 5 has a self-referential quality: the collaboration process used to build the research program becomes the object of study. The research program studies itself.
| File | Status | Description |
|---|---|---|
paper.md |
✅ Complete | Full paper — 13 chapters, HARCT framework |
| Chapter | Title |
|---|---|
| 1 | Introduction — motivation, research questions, contributions |
| 2 | Background — Human-AI collaboration, context windows, knowledge preservation |
| 3 | Research Environment — participants, repositories, documentation systems |
| 4 | Human-AI Research Continuity Theory (HARCT) — central framework |
| 5 | Goal Preservation Theory — goals > information; Future-Self Collaboration |
| 6 | Cognitive Architecture — memory asymmetry, labor division, repository as cognitive infrastructure |
| 7 | Reconstruction Cost Theory — documentation as economic variable |
| 8 | Externalized Memory and AI-to-AI Handover |
| 9 | Token Constraints and Context Loss |
| 10 | Building a Long-Term Research Program |
| 11 | Case Studies — real-world observations from the Minor Thesis program |
| 12 | Discussion — HARCT unified framework, testable predictions, failure modes, boundary conditions |
| 13 | Conclusion — summary, future directions, formal HARCT variables |
Long-term Human-AI collaboration is fundamentally a problem of research continuity rather than intelligence alone.
| Component | Claim |
|---|---|
| Goal Preservation Theory | Goals are more fundamental than information; goals enable reconstruction |
| Cognitive Architecture | Human goal memory + AI reasoning + artifact memory = distributed continuity system |
| Reconstruction Cost Theory | Documentation quality is an economic variable determining collaboration efficiency |
| Externalized Memory | Context is preserved outside active participants through persistent artifacts |
| Concept | Definition |
|---|---|
| Externalized Memory | Research context stored outside human memory and AI context windows |
| Distributed Research Memory | Memory distributed across humans, AI systems, repositories, and artifacts |
| Future-Self Collaboration | Collaboration between temporally displaced versions of the same researcher |
| Reconstruction Cost | Effort required to restore sufficient context for productive research |
| Memory Repository | Any artifact capable of preserving and restoring research context |
| AI-to-AI Handover | Context transfer mechanism enabling AI system transitions |
| Concept | Novelty |
|---|---|
| HARCT (central framework) | High — long-term multi-AI collaboration as continuity problem |
| Goal Preservation Theory | High — goals > information in Human-AI collaboration |
| Future-Self Collaboration | High — documentation as collaboration with a temporally displaced self |
| Reconstruction Cost Theory | High — documentation quality as explicit economic variable |
| Cognitive Labor Division | Medium — overlaps with Distributed Cognition |
| Prediction | Claim |
|---|---|
| P1 | Better repository infrastructure → higher continuity |
| P2 | Higher documentation quality → lower reconstruction cost |
| P3 | Structured handover artifacts → less context loss at AI transitions |
| P4 | Future-oriented artifacts → faster context recovery |
| P5 | Goal preservation → projects survive longer interruptions |
| P6 | Reconstruction cost scales with research duration |
| P7 | Artifact-centered knowledge → higher AI replaceability |
| P8 | Stable collaboration protocols → reduced context re-establishment cost |
| Item | State |
|---|---|
Main paper (paper.md) |
✅ Complete — 13 chapters |
GitHub Pages (index.html) |
🔧 Pending setup |
README (README.md) |
✅ This file |
This repository is part of the Structure Recognition Research Program.
| Repository | Role |
|---|---|
| Research-Portfolio | Program Hub |
| 1KMapStructureInvariance | Empirical Case Study 1 |
| 2SymmetricBooleanFunctionMinorThesis | Empirical Case Study 2 |
| 3VariableRearrangementInvarianceMinorThesis | Empirical Case Study 3 |
| 4StructureRecognitionTheory | Theoretical Hub |
| 5HumanAIResearchCollaboration | Methodological Meta-layer ← You are here |
Master Handover Document: MasterHandoverDocument.md
Choi Jonghun Independent researcher · Graduate of Inha Technical College