Risky collaborators: qualitative research ethics in an era of algorithmic co-presence

Existing ethical discussions of artificial intelligence (AI) often position it as a technical instrument, reducing ethics to data security, privacy, and intellectual property while sidelining its participation in the production of qualitative knowledge. While important, these perspectives are insufficient for addressing the conditions through which AI reshapes inquiry. In response, we develop a three-plane framework for examining these conditions across material–structural, epistemic–distributive, and relational–performative dimensions. Together, the planes form a recursive circuit linking the infrastructures that sustain AI, the uneven knowledge records from which systems learn, and the research encounters through which their outputs are taken up and carried forward. We argue that epistemic erasure under AI-assisted inquiry can therefore compound across successive encounters, with researchers participating in the loops they seek to examine. We reorient qualitative research ethics toward response-ability as a situated and distributed practice of recognizing, contesting, and interrupting these recursive conditions from within.