Human visual explanations mitigate bias in AI-based assessment of surgeon skills.
Human visual explanations mitigate bias in AI-based assessment of surgeon skills.
复制标题
DOI：
10.1038/s41746-023-00766-2
复制
发表时间：
2023-03-30
影响因子：
15.2
通讯作者：
中科院分区：
文献类型：
--
作者：
人工智能（AI）系统现在可以通过术中手术活动的视频可靠地评估外科医生的技能。随着这些系统为未来的高风险决策提供信息，例如是否授予外科医生资格并授予他们为患者手术的特权，公平对待所有外科医生至关重要。然而，手术AI系统是否对外科医生亚组群表现出偏倚仍然是一个悬而未决的问题，如果是这样，这种偏倚是否可以减轻。在这里，我们检查并减轻了一系列手术AI系统-SAIS-在三家地理位置不同的医院（美国和欧盟）的机器人手术视频上部署的偏见。我们发现，SAIS表现出技能不足的偏见，错误地降低手术性能，和过度的偏见，错误地提高手术性能，在不同的利率在外科医生的子队列。为了减轻这种偏见，我们利用了一种策略-TWIX-它教导AI系统为其技能评估提供视觉解释，否则将由人类专家提供。我们发现，尽管基线策略不一致地减轻了算法偏差，但TWIX可以有效地减轻技能不足和技能过度的偏差，同时提高这些AI系统在医院中的性能。我们发现，这些发现延续到我们今天评估医学生技能的培训环境中。我们的研究是最终实施人工智能增强的全球外科医生认证计划的关键先决条件，确保所有外科医生都得到公平对待。
Artificial intelligence (AI) systems can now reliably assess surgeon skills through videos of intraoperative surgical activity. With such systems informing future high-stakes decisions such as whether to credential surgeons and grant them the privilege to operate on patients, it is critical that they treat all surgeons fairly. However, it remains an open question whether surgical AI systems exhibit bias against surgeon sub-cohorts, and, if so, whether such bias can be mitigated. Here, we examine and mitigate the bias exhibited by a family of surgical AI systems—SAIS—deployed on videos of robotic surgeries from three geographically-diverse hospitals (USA and EU). We show that SAIS exhibits an underskilling bias, erroneously downgrading surgical performance, and an overskilling bias, erroneously upgrading surgical performance, at different rates across surgeon sub-cohorts. To mitigate such bias, we leverage a strategy —TWIX—which teaches an AI system to provide a visual explanation for its skill assessment that otherwise would have been provided by human experts. We show that whereas baseline strategies inconsistently mitigate algorithmic bias, TWIX can effectively mitigate the underskilling and overskilling bias while simultaneously improving the performance of these AI systems across hospitals. We discovered that these findings carry over to the training environment where we assess medical students’ skills today. Our study is a critical prerequisite to the eventual implementation of AI-augmented global surgeon credentialing programs, ensuring that all surgeons are treated fairly.
登录
查看更多内容
影响因子：
13.8
作者：
Khalid, Shuja;Goldenberg, Mitchell;Rudzicz, Frank
通讯作者：
Rudzicz, Frank
DOI：
10.1016/j.juro.2017.07.081
发表时间：
2018-01
期刊：
The Journal of urology
影响因子：
--
作者：
Hung AJ;Chen J;Jarc A;Hatcher D;Djaladat H;Gill IS
通讯作者：
Gill IS
影响因子：
16.6
作者：
Kiyasseh D;Zhu T;Clifton D
通讯作者：
Clifton D
影响因子：
6.6
作者：
Hung, Andrew J.;Zehnder, Pascal;Desai, Mihir M.
通讯作者：
Desai, Mihir M.
影响因子：
9
作者：
Gallagher, AG;Ritter, EM;Satava, RM
通讯作者：
Satava, RM