Patients' experiences and attitudes toward the use of real-time artificial intelligence-assisted colonoscopy

There is limited knowledge about how patients experience artificial intelligence (AI) in healthcare. Concerns regarding safety and quality could be a barrier between patients and the healthcare system when implementing AI. It is crucial to understand patients' attitudes, and how they may react to AI being implemented when aiming to preserve their trust. We investigated patients' attitudes toward AI-assisted colonoscopy to gain insights that can be used when introducing this new technology to future patients.

This is a qualitative study based on semi-structured focus group interviews. Habile, Danish-speaking adults who had received an AI-assisted colonoscopy were invited. We conducted six focus group interviews, with 20 participants in total. All interviews were recorded and transcribed. Data were analysed using thematic analysis.

Patients described a sense of vulnerability in relation to undergoing a colonoscopy. Three main themes were identified. The first theme is 'Empathy and professionalism-trust in AI-assistance in colonoscopy depends on human factors'. The second and the third themes are 'Information about AI should be proportional to the consequences', and 'Trying to make sense of AI during colonoscopy-balancing curiosity, irrelevance, and vulnerability'.

Trust in AI-assistance is dependent on trust in the clinicians. While AI has many strengths, it cannot provide empathy, which is essential for patients in distressing moments, such as during a colonoscopy. AI should be considered a supportive tool, while the endoscopist remains responsible for clinical decisions and patient outcomes. If written information about AI is provided, it should be informative but concise, avoiding details that generate more confusion than clarity.