Artificial Intelligence and the Clinical Gaze: Visual Practices of AI‐Assisted Colonoscopy

TL;DR

This study explores how AI systems are changing how doctors visually examine the colon during colonoscopies.

Contribution

The paper introduces three modes of AI integration in colonoscopy—technology-driven, experience-driven, and interrupted—highlighting how they reshape medical visual practices.

Findings

AI integration in colonoscopy involves real-time interaction between doctors and algorithmic markings.

The clinical gaze is reconfigured through embodied routines and technological infrastructures.

AI does not replace the clinical gaze but alters its conditions of enactment and introduces new uncertainties.

Abstract

This article examines how commercial artificial intelligence (AI) systems are integrated into the visual practice of colonoscopy and how they reshape the clinical gaze as a sociotechnical and situated mode of perception. Based on ethnographic observations of colonoscopies and in‐depth interviews with gastroenterologists, this study analyses the real‐time use of AI‐based pattern recognition systems during diagnostic procedures. Unlike retrospective image analysis in radiology, AI in colonoscopy operates in vivo and in real time, requiring practitioners to engage with algorithmic markings within the unfolding process of examination. The clinical gaze emerges here as a form of professional vision constituted through embodied routines, tacit knowledge, technological infrastructures and institutionalised practices. This article identifies three modes of integrating AI—technology‐driven, experience‐driven and interrupted—that reveal the contingency and multiplicity of human–machine relations in clinical imaging work. Rather than displacing the clinical gaze, AI alters the conditions under which it can be enacted, bringing with it shifting forms of visual selectivity, epistemic authority and new uncertainties. This study contributes to sociological debates on algorithmic medicine and digital automation by showing how AI systems intervene in the practical organisation of medical seeing, highlighting the tensions and adjustments through which contemporary visual practices in healthcare are reconfigured.