Beyond Efficiency: A Triadic Framework for AI-Assisted Qualitative Inquiry Honoring Researcher Sovereignty

• In Event: Critical and Culturally Conscious Engagement with Generative AI in Qualitative Research

Abstract

Objectives
In this paper, I present a framework for Artificial-Intelligence (AI) assisted qualitative inquiry, the Triadic AI Engagement Framework (TAF). Using de/colonizing ontoepistemologies (Bhattacharya, 2022; Smith, 2012), I demonstrate how this framework enables qualitative researchers to maintain sovereignty while engaging AI that move beyond time-saving or productivity gains. My objectives: (a) explore how current literature's efficiency obsession erases the in-depth care and cultural situatedness essential to qualitative inquiry, (b) present my triadic framework grounded in Anzaldúa's nepantla and borderlands sensibilities, (c) demonstrate practical applications that preserve researcher sovereignty and honor diverse epistemologies, and (d) share how de/colonizing ontoepistemologies can neutralize multiple forms of algorithmic biases.

Perspectives
Using TAF, I draw on three theoretical foundations. First, de/colonial scholarship (Bhattacharya, 2022; Smith, 2012) informs my understanding of how technological systems embed colonial logics and interconnected algorithmic biases that must be actively resisted. Second, Anzaldúa's (1987, 2015) nepantla and borderlands theory grounds navigating liminal spaces between human interpretation and algorithmic processing. Third, Robbins' (2023) last mile concept highlights the necessity for quality control at the intersection of human and algorithmic processes, as represented in AI outputs.
Intersecting these perspectives, TAF identifies three fluid, interconnected roles that shape human-AI interactions: World Builders, Bridge Builders, and Last Mile Workers. World Builders design AI systems, embedding western epistemologies into algorithms and training data. Bridge Builders embody Anzaldúa's (2015) nepantleras—threshold people who translate between technological capabilities and cultural contexts while refusing exclusive allegiance to either. They navigate the nepantla of AI engagement, maintaining what Anzaldúa (2015) calls "la facultad" to read multiple realities simultaneously. Last Mile Workers operationalize Robbins' (2023) concept, providing human discernment, quality control on the output provided for accuracy, nuance, depth, and context.

Methods
Through a combination of theoretical analysis, discourse analysis of AI-centered conversations in online forums, and interrogative experimental practice, and reviewing the literature in AI-assisted work in higher education and qualitative research, I developed TAF. To move beyond productivity metrics, I engaged in iterative dialogue with AI systems to test whether they could honor thoughtful, culturally situated approaches. My process explored how de/colonizing approaches might interrupt algorithmic biases while preserving the slow, careful work of interpretation, analysis, and representation in qualitative research. As a conceptual paper, I will demonstrate TAF's foundational and practical applications through examples of: (a) using AI as a de/colonial memory keeper (b) engaging AI as a theorizing and an intellectual sparring partner, and (c) showing how researchers can move fluidly between World Builder, Bridge Builder, and Last Mile Worker roles based on their needs.

Significance
Utilizing TAF and aligning with AERA's 2026 call for "constructing new visions," qualitative researchers can engage with AI-assisted qualitative research that preserves the in-depth care and cultural situatedness, while maintaining the criticality of researcher sovereignty and interpretive authority. Crucially, these applications require no participant data uploads and explicitly reject efficiency as a primary value. While not all qualitative researchers are required to use AI, but those who are curious and cautious can use this paper as a possibility model.

Author

• Kakali Bhattacharya, University of Florida