Advances in artificial intelligence (AI) have radically transformed healthcare for both patients and healthcare providers, from early detection and diagnosis to clinical decision-making to delivery of care. AI systems have the potential to reduce costs of healthcare and facilitate more efficient, precise, and personalized care. Yet there is rising concern for the dark side of AI systems, such as concerns of low-usability, vulnerability, “black-box”, bias, prejudice, discrimination, and privacy. This has spurred the field of Responsible AI, in an effort to build high-usability, robustness, fairness, accountability, transparency, and privacy into AI systems. These principles and values are of paramount importance in high-stake fields such as healthcare. The increasing emphasis on ethical and responsible AI points to an urgent need to consider normative, regulatory, and ethical challenges facing AI systems in healthcare, and delineate approaches to minimize the potential harms in these systems.
The goal of this Research Topic is to understand the challenges facing AI systems for healthcare in terms of fairness, accountability, explainability, and privacy, and best practices to resolve these issues. We are particularly interested in process challenges for developing, designing, integrating, and evaluating clinical AI tools. These might include technical difficulties in machine learning (ML) processes, such as data collection and cleaning, model development and deployment, as well as practical difficulties in actual implementation in clinical settings and user interaction with AI systems. We seek to highlight the principles and values of Responsible AI in building healthcare systems, which can facilitate more equitable algorithms, compassionate design, and sustainable solutions.
Relevant submissions for this Research Topic include, but are not limited to, the following:
• UX research on user interaction with digital health technologies to inform ethical, responsible, and explainable AI in healthcare
• Empirical and conceptual research on how to balance needs for utility, fairness, privacy, and safety.
• Empirical studies examining clinician and patient perceptions of AI systems in healthcare, and the impact of such systems on decision making, work practices, identities, etc.
• Responsible AI enabled human-AI collaboration/teaming where human and AI together achieves better outcomes
• Studies of legislative, professional, and regulatory frameworks for responsible AI in healthcare, such as algorithmic impact assessment, and implementation of such frameworks
• Novel human-centered AI system that aims to incorporate human-centered design principles
• Cross-cultural comparisons and assessments of ethical guidelines and frameworks governing AI systems in healthcare
• Technical studies that detect and/or mitigate algorithmic bias in clinical AI systems
• Algorithms and toolkits to build explanatory, transparent, and privacy-preserving machine learning models for healthcare technologies
• Novel data sets and data capture methods to inform Responsible AI in healthcare
Advances in artificial intelligence (AI) have radically transformed healthcare for both patients and healthcare providers, from early detection and diagnosis to clinical decision-making to delivery of care. AI systems have the potential to reduce costs of healthcare and facilitate more efficient, precise, and personalized care. Yet there is rising concern for the dark side of AI systems, such as concerns of low-usability, vulnerability, “black-box”, bias, prejudice, discrimination, and privacy. This has spurred the field of Responsible AI, in an effort to build high-usability, robustness, fairness, accountability, transparency, and privacy into AI systems. These principles and values are of paramount importance in high-stake fields such as healthcare. The increasing emphasis on ethical and responsible AI points to an urgent need to consider normative, regulatory, and ethical challenges facing AI systems in healthcare, and delineate approaches to minimize the potential harms in these systems.
The goal of this Research Topic is to understand the challenges facing AI systems for healthcare in terms of fairness, accountability, explainability, and privacy, and best practices to resolve these issues. We are particularly interested in process challenges for developing, designing, integrating, and evaluating clinical AI tools. These might include technical difficulties in machine learning (ML) processes, such as data collection and cleaning, model development and deployment, as well as practical difficulties in actual implementation in clinical settings and user interaction with AI systems. We seek to highlight the principles and values of Responsible AI in building healthcare systems, which can facilitate more equitable algorithms, compassionate design, and sustainable solutions.
Relevant submissions for this Research Topic include, but are not limited to, the following:
• UX research on user interaction with digital health technologies to inform ethical, responsible, and explainable AI in healthcare
• Empirical and conceptual research on how to balance needs for utility, fairness, privacy, and safety.
• Empirical studies examining clinician and patient perceptions of AI systems in healthcare, and the impact of such systems on decision making, work practices, identities, etc.
• Responsible AI enabled human-AI collaboration/teaming where human and AI together achieves better outcomes
• Studies of legislative, professional, and regulatory frameworks for responsible AI in healthcare, such as algorithmic impact assessment, and implementation of such frameworks
• Novel human-centered AI system that aims to incorporate human-centered design principles
• Cross-cultural comparisons and assessments of ethical guidelines and frameworks governing AI systems in healthcare
• Technical studies that detect and/or mitigate algorithmic bias in clinical AI systems
• Algorithms and toolkits to build explanatory, transparent, and privacy-preserving machine learning models for healthcare technologies
• Novel data sets and data capture methods to inform Responsible AI in healthcare