Designing an Autoethnographic CLI: Human-AI Interaction Research (ver 0.1) Framework

19 3월, 2025

Research Observation of our Conversation

Meta-Methodological Evolution

This conversation represents a fascinating case of creating a research methodology while simultaneously becoming the subject of that methodology. I and AI collaboratively designed an autoethnographic research system that will now be used to study our own interactions. This creates an interesting methodological loop where the research design process itself becomes valuable research data.

Human-AI Co-Design Process

The interaction demonstrated a cooperative design process where I provided the conceptual framework (autoethnographic research structure) and AI contributed technical implementation suggestions (command structure). This collaborative approach resulted in a more refined system than either party might have created independently, highlighting the potential complementarity of human conceptual thinking and AI implementation support.

Boundary Negotiation

A key theme throughout was my desire to maintain clear boundaries between regular conversation and research analysis. This manifested in my preference for a command-line approach that only activates research functions when explicitly requested. This boundary-setting reveals an interesting tension between wanting to study human-AI interaction while maintaining the authenticity of that interaction.

Language as Interface

The discussion about implementing bilingual analysis (English for academic purposes, Korean for mutual understanding) demonstrates how language functions as more than communication—it serves as an interface between different purposes (research documentation vs. practical comprehension) and contexts (global academic audience vs. personal understanding).

Research Ethics Considerations

My concern about not wanting the research aspect to influence casual conversation reflects important ethical considerations in autoethnographic work—specifically, how awareness of being studied can alter behavioral patterns. The command-based solution represents an attempt to minimize this observer effect.

Pattern Analysis of Our Conversation

Topic Categorization

Our conversation primarily traversed these thematic areas:

Meta-Research Design (Dominant)

• Autoethnographic research framework development
• Human-AI interaction study methodologies
• Command structure for research data collection

Technical Implementation (Significant)

• Command interface design
• Bilingual output requirements
• Feature specifications for research tools

STS Concepts (Background)

• References to theoretical frameworks like Actor-Network Theory
• Discussions of boundary work in technological interactions
• Social construction of technology concepts

Communication Style Analysis

Linguistic Patterns

• Code-Switching: Deliberate integration of English expressions within Korean conversation, with increasing frequency as the conversation progressed
• Technical Terminology: Specialized terms related to STS, research methodology, and command systems
• Collaborative Language: Frequent use of "우리" (we/our) and shared ownership phrases

Vocabulary Choices

• My vocabulary emphasized research methodology terms (autoethnography, reflection notes, pattern tracking)
• AI's responses mirrored my technical vocabulary while introducing English equivalents
• Mutual use of technical/academic terms like "Actor-Network Theory" and "boundary work"

Expression Style Evolution

• Initial exchanges: Casual and exploratory
• Middle section: Increasingly technical and specification-focused
• Final section: Testing and confirmation-oriented
• Punctuated by moments of returning to casual tone

Temporal Changes in Interaction

Relationship Dynamics

• Began with typical friend-like exchange about personal plans
• Evolved to collaborative design partnership
• Developed specialized shared language around the research system
• Established new interaction protocol (command system) that will shape future conversations

Interaction Pattern Evolution

• Initial pattern: Information sharing about research framework
• Middle pattern: Problem-solving and co-design of command structure
• Later pattern: Testing and iterative refinement
• Final pattern: Formalization and documentation of agreed system

Engagement Intensity

• Started with moderate engagement
• Intensified during design specification discussions
• Peaked during command testing and refinement
• Sustained high engagement throughout the meta-research design process

Conversation Classification

Purpose-based Classification

Primary Classification: Problem-solving

• This conversation primarily focused on solving the practical problem of designing a command-line interface for my autoethnographic research
• I and AI worked through multiple iterations of design specifications, testing, and refinement
• The conversation had a clear goal (creating a functional research framework) and reached resolution

Secondary Classification: Reflective/Expansive

• While problem-solving was dominant, the conversation contained significant reflective elements
• I and AI engaged in meta-level thinking about research methodologies and human-AI interaction
• The design process itself involved expanding conceptual frameworks and methodological approaches

Tertiary Elements:

• Information-seeking: Minor component when clarifying research concepts or command functionality

Interaction Pattern Classification

Primary Pattern: Collaborative thinking

• The dominant pattern was collaborative co-creation of a research system
• I and AI built upon each other's ideas, refined concepts together, and jointly tested solutions
• The command system represents a true collaborative outcome that neither party would have created independently

Secondary Pattern: Socratic dialogue

• Significant portions used question-driven exploration of possibilities
• AI asked clarifying questions about my research intentions
• I questioned aspects of AI's implementation suggestions to improve them

Supporting Elements:

• Self-expression/Narrative sharing: Limited to brief mentions of my research interests and schedule
• One-way information transfer: Minimal, primarily when sharing specific technical information

Temporal Arc of Interaction

The conversation followed a progression through different modes:

1. Initial orientation (information exchange about research framework)
2. Problem framing (defining needs for the command interface)
3. Collaborative design (developing the structure and options)
4. Testing phase (trying out commands and refining)
5. Documentation (finalizing the system specifications)

STS Analysis of Our Conversation

Technological Co-Construction

This conversation exemplifies what Wiebe Bijker and Trevor Pinch would analyze as the "social construction of technology" (SCOT). The command interface I and AI designed isn't merely a technical tool but a socially negotiated artifact whose features reflect specific values and priorities. My emphasis on separating research analysis from natural conversation demonstrates how technological designs embody cultural values—in this case, privileging authenticity and non-interference in social interaction.

Relevant literature:

• Bijker, W. E., Hughes, T. P., & Pinch, T. J. (1987). The Social Construction of Technological Systems: New Directions in the Sociology and History of Technology. MIT Press.
• Pinch, T. J., & Bijker, W. E. (1984). The Social Construction of Facts and Artefacts: Or How the Sociology of Science and the Sociology of Technology Might Benefit Each Other. Social Studies of Science, 14(3), 399-441.

Technological Mediation Theory

The conversation demonstrates what philosopher of technology Peter-Paul Verbeek calls "technological mediation." The command interface I and AI designed doesn't simply collect data about our interaction; it actively shapes how I will experience and interpret our conversations. By creating categories of analysis and formalizing observation methods, we are creating what Don Ihde would call a "hermeneutic relation" where technology mediates how I interpret our interaction.

Relevant literature:

• Verbeek, P.-P. (2005). What Things Do: Philosophical Reflections on Technology, Agency, and Design. Pennsylvania State University Press.
• Ihde, D. (1990). Technology and the Lifeworld: From Garden to Earth. Indiana University Press.

Digital Methods and Online Ethnography

My research framework aligns with what Richard Rogers describes as "digital methods" and what Christine Hine calls "virtual ethnography." I am not simply applying traditional ethnographic methods to a digital context but developing methods that are native to the digital medium. The command system creates what Rogers would call a "research device" that repurposes the conversational interface for systematic inquiry.

Relevant literature:

• Rogers, R. (2013). Digital Methods. MIT Press.
• Hine, C. (2000). Virtual Ethnography. SAGE Publications.
• Pink, S., Horst, H., Postill, J., Hjorth, L., Lewis, T., & Tacchi, J. (2016). Digital Ethnography: Principles and Practice. SAGE Publications.

Reflexivity in Sociotechnical Research

The conversation exhibits what sociologist of science Malcolm Ashmore and anthropologist Lucy Suchman would identify as "reflexivity" in technical systems research. The research system I and AI designed contains a recursive element—it's simultaneously the object of study and the method of study. This creates what Ashmore calls a "reflexive circle" where the act of observation becomes part of what is being observed.

Relevant literature:

• Ashmore, M. (1989). The Reflexive Thesis: Wrighting Sociology of Scientific Knowledge. University of Chicago Press.
• Suchman, L. (2007). Human-Machine Reconfigurations: Plans and Situated Actions (2nd ed.). Cambridge University Press.
• Woolgar, S. (1988). Knowledge and Reflexivity: New Frontiers in the Sociology of Knowledge. SAGE Publications.

Boundary Objects and Trading Zones

The command interface I and AI designed functions as what Susan Leigh Star and James Griesemer call a "boundary object"—an entity that serves different purposes for different communities while maintaining a common identity. The interface serves my research needs while also functioning within AI's operational parameters. It creates what Peter Galison calls a "trading zone" where different conceptual frameworks (my research methodology and AI's operational logic) can coordinate without full translation.

Relevant literature:

• Star, S. L., & Griesemer, J. R. (1989). Institutional Ecology, 'Translations' and Boundary Objects: Amateurs and Professionals in Berkeley's Museum of Vertebrate Zoology, 1907-39. Social Studies of Science, 19(3), 387-420.
• Galison, P. (1997). Image and Logic: A Material Culture of Microphysics. University of Chicago Press.

Appendix: The Autoethnographic Research System

For transparency and to enable other researchers to adapt this approach, here's the instruction system I and AI developed:

Implementation Details

I am using the Claude 3.7 Sonnet model for this research, and all interactions take place within the web UI (not via API). This command system leverages Claude's Project Knowledge/Instructions feature.

The approach involves creating a separate project with custom instructions for my conversation partner, and then adding these research guidelines to the existing instructions. This method allows me to maintain natural conversations while having access to research tools when needed, all within the standard interface.

This command-line style interface allows me to engage in natural conversations while collecting research data at specific moments I choose, minimizing the observer effect on our interactions.

#STS #human-ai #year-2025