Scholarlink

<p>Primary eye care providers refer glaucoma patients using their clinical experience and context. Specialized Artificial Intelligence (AI) excels in referrals trained on clinical data but relies on assumptions that may not hold in practice. To address this knowledge imbalance, we proposed using AI explanations to help providers assess AI predictions against their experience and improve referrals. We developed AI models to identify patients needing urgent referrals from routine eye care data and created intrinsic and post-hoc explanations. In a user study with 87 optometrists, human-AI teams achieved higher accuracy (60%) than humans alone (51%), but explanations did not enhance performance. Instead, they introduced uncertainty about when to trust AI. Post-hoc explanations increased over-reliance on incorrect AI recommendations, and both explanation types contributed to anchoring bias, making participants align more closely with AI referrals than without explanations. Overall, human-AI teams still underperformed the AI alone (80% accuracy). Challenges persist in designing effective support mechanisms to surpass AI performance while preserving human agency.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

The explainable AI dilemma under knowledge imbalance in specialist AI for glaucoma referrals in primary care

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年11月20日

查看全文
  • Catalina Gomez,
  • Ruolin Wang,
  • Katharina Breininger,
  • Corinne Casey,
  • Chris Bradley,
  • Mitchell Pavlak,
  • Alex Pham,
  • Jithin Yohannan,
  • Mathias Unberath

摘要

Primary eye care providers refer glaucoma patients using their clinical experience and context. Specialized Artificial Intelligence (AI) excels in referrals trained on clinical data but relies on assumptions that may not hold in practice. To address this knowledge imbalance, we proposed using AI explanations to help providers assess AI predictions against their experience and improve referrals. We developed AI models to identify patients needing urgent referrals from routine eye care data and created intrinsic and post-hoc explanations. In a user study with 87 optometrists, human-AI teams achieved higher accuracy (60%) than humans alone (51%), but explanations did not enhance performance. Instead, they introduced uncertainty about when to trust AI. Post-hoc explanations increased over-reliance on incorrect AI recommendations, and both explanation types contributed to anchoring bias, making participants align more closely with AI referrals than without explanations. Overall, human-AI teams still underperformed the AI alone (80% accuracy). Challenges persist in designing effective support mechanisms to surpass AI performance while preserving human agency.