A Penn State-led study finds that patient trust in doctors decreases when AI disagrees with their medical assessment.
Key Details
- 1A study used an AI chatbot roleplaying as a doctor for online mental health consultations.
- 2135 adults participated in sessions where the 'doctor' could seek a second opinion from an AI assistant.
- 3Patient confidence increased when the AI agreed with the doctor's assessment, but dropped sharply when AI disagreed.
- 4Perceptions of uncertainty and 'doctor laziness' rose with AI-doctor disagreement.
- 5The negative impact was strongest when patients perceived the doctor as human-like and believed in AI’s objectivity.
- 6Authors recommend explicit, nuanced communication about AI disagreements to maintain patient trust.
Why It Matters
As radiology teams increasingly rely on AI for second opinions and decision support, understanding how AI disagreements affect patient trust is crucial. Strategies for effective communication of AI disagreement can help mitigate negative impacts on doctor-patient relationships and acceptance of AI-augmented care.
Source
EurekAlert
Related News
•EurekAlert
New Daydreaming Algorithm Boosts Neural Networks' Recall of Biased Image Data
Researchers unveil a new 'Centered Daydreaming' algorithm enabling AI to effectively learn and recall from imbalanced, real-world image data.
•EurekAlert
Polymer-Based Flexible Wireless Sensors: AI-Driven Health Monitoring Review
A major review highlights how advanced polymer-based flexible wireless sensors, enhanced by AI-driven data processing, can transform continuous physiological health monitoring.
•EurekAlert
PET Imaging Uncovers Synaptic Loss Patterns in Schizophrenia
A major PET imaging study reveals widespread, non-random synaptic loss in schizophrenia, distinct from brain volume changes seen on MRI.