The Identification and Prevention Strategies of Liability Risks in the Application of Artificial Intelligence in the Public Sector
Artificial Intelligence(AI)technology and service presents a series of ethical risks of uncertain modernity in their development,deployment and application.It profoundly affects digital governance and has become a key topic in the field.Because of its negative impact on social value and the predictability of social trust,liability risk is an important ethi-cal risk problem that government departments cannot avoid in digital governance.Existing academic research focuses on the identification and control of AI ethical risks at the macro level,but lacks in-depth exploration of the identification and pre-vention strategies for specific ethical risk patterns such as liability risk.The reciprocal relationship between technology and institutions offers a sound framework for typology research.Taking the conflict between instrumental rationality and value ra-tionality as the logical starting point,the identification framework of AI responsibility risk is constructed.Responsibility risk is divided into four patterns:disordered attribution of responsibility,cognitive dissonance of responsibility,misbehavior of responsibility and imbalance of responsibility value.The research is centered on four types of AI responsibility risk pat-terns:distorted responsibility attribution,cognitive dissonance in responsibility,malfeasance in responsibility behavior,and imparity in responsibility value.We will build a"technology-organization-institution"(TOI)preventive mechanism.It is inevitable to enhance the predictability of risks and the resilience of risk management in government department,so as to realize safe and responsible digital governance.
万方数据
维普
