[.] we found a persistent bias in which human errors are seen as forgivable, but machine mistakes are not.
Philip Samell & Sigve Angell Spikseth
No second chance for the machine
The increasing integration of artificial intelligence (AI) in healthcare holds significant promise, but individuals often prefer a human touch in decision-making, dislike being reduced to a percentage by algorithms and value dignity and respect which may be compromised by AI systems.
Through an experimental study, our master’s thesis at the Norwegian School of Economics (NHH), part of the DIG program, investigates the question of how perceptions of trust influence the choice between an AI system and a general practitioner (GP) in medical consultation, and if a misdiagnosis changes these perceptions.
The Role of Trust in Healthcare Decisions
The nature of healthcare such as uncertainty, information asymmetries and high stakes makes trust is pivotal in healthcare. We measured trust through three key dimensions: ability, benevolence, and integrity. Our findings revealed that AI systems lacked elements of benevolence and integrity which was associated with general practitioners. These human-like qualities are critical in emotionally significant contexts, where empathy and personalized care are central.
Algorithm Aversion and Patient Preferences
Despite AI’s higher diagnostic accuracy in complex cases, many participants expressed hesitance toward relying on AI in high-complexity symptom scenarios. This is related to algorithm aversion, where we found a persistent bias in which human errors are seen as forgivable, but machine mistakes are not. Our study found that patients experienced greater regret when misdiagnosed by AI compared to a GP, highlighting the stronger sense of accountability towards AI systems. Interestingly, participants with prior experience of AI systems were more likely to choose AI over a GP which underscores the need for greater transparency and education about AI capabilities to reduce skepticism and build trust.
Creating a Balance Between Innovation and Empathy
Our findings emphasize the importance of combining AI’s diagnostic strengths with the empathetic care general practitioners provide. To integrate AI effectively, healthcare systems must prioritize transparency, accountability, and patient-centered innovations. As healthcare evolves, fostering trust in AI is not just a technical challenge—it’s an emotional one. By addressing concerns about accountability and embracing both innovation and empathy, we