Combining the computational power of artificial intelligence (AI) with the coercive power of the state strikes many people as deeply troubling. Governments are currently using or developing AI applications that seem to increase racial bias, invade people's privacy, prevent people from understanding the reasoning behind decisions that adversely affect them, and ensure that no one could justly be held responsible for the results of using them. This has led a growing number of writers in the AI ethics literature to argue that these unsettling features of contemporary AI technology make it wrong for states to use AI in these ways even if doing so would otherwise be socially beneficial.In this timely and provocative book, David Boonin challenges this emerging consensus. Focusing on four widely-discussed case studies, Boonin defends four striking and controversial claims: that causing police deployment decisions to be more racially biased does not make it wrong to use predictive policing algorithms if using them reduces crime; that using CCTV-based mass public surveillance to reduce crime does not violate people's right to privacy; that using opaque recidivism prediction algorithms to help officials in the criminal justice system make decisions about bail, sentencing, parole, and probation does not violate people's right to get an explanation of the steps in the reasoning that led to the decisions in their cases; and that developing highly autonomous lethal weapon systems that outperform human soldiers would not create a morally objectionable responsibility gap in which no one could justly be held responsible for a weapon system's behavior. Boonin also argues that each of these conclusions can be successfully extended to a variety of other important cases. Written in a clear and accessible style, Artificial Intelligence, Ethics, and the State will be of interest to students and scholars in a variety of fields as well as to anyone concerned about the growing use of AI by governments around the world.

Combining the computational power of artificial intelligence (AI) with the coercive power of the state strikes many people as deeply troubling. Governments are currently using or developing AI applications that seem to increase racial bias, invade people's privacy, prevent people from understanding the reasoning behind decisions that adversely affect them, and ensure that no one could justly be held responsible for the results of using them. This has led a growing number of writers in the AI ethics literature to argue that these unsettling features of contemporary AI technology make it wrong for states to use AI in these ways even if doing so would otherwise be socially beneficial.In this timely and provocative book, David Boonin challenges this emerging consensus. Focusing on four widely-discussed case studies, Boonin defends four striking and controversial claims: that causing police deployment decisions to be more racially biased does not make it wrong to use predictive policing algorithms if using them reduces crime; that using CCTV-based mass public surveillance to reduce crime does not violate people's right to privacy; that using opaque recidivism prediction algorithms to help officials in the criminal justice system make decisions about bail, sentencing, parole, and probation does not violate people's right to get an explanation of the steps in the reasoning that led to the decisions in their cases; and that developing highly autonomous lethal weapon systems that outperform human soldiers would not create a morally objectionable responsibility gap in which no one could justly be held responsible for a weapon system's behavior. Boonin also argues that each of these conclusions can be successfully extended to a variety of other important cases. Written in a clear and accessible style, Artificial Intelligence, Ethics, and the State will be of interest to students and scholars in a variety of fields as well as to anyone concerned about the growing use of AI by governments around the world.

Combining the computational power of artificial intelligence (AI) with the coercive power of the state strikes many people as deeply troubling. Governments are currently using or developing AI applications that seem to increase racial bias, invade people's privacy, prevent people from understanding the reasoning behind decisions that adversely affect them, and ensure that no one could justly be held responsible for the results of using them. This has led a growing number of writers in the AI ethics literature to argue that these unsettling features of contemporary AI technology make it wrong for states to use AI in these ways even if doing so would otherwise be socially beneficial.In this timely and provocative book, David Boonin challenges this emerging consensus. Focusing on four widely-discussed case studies, Boonin defends four striking and controversial claims: that causing police deployment decisions to be more racially biased does not make it wrong to use predictive policing algorithms if using them reduces crime; that using CCTV-based mass public surveillance to reduce crime does not violate people's right to privacy; that using opaque recidivism prediction algorithms to help officials in the criminal justice system make decisions about bail, sentencing, parole, and probation does not violate people's right to get an explanation of the steps in the reasoning that led to the decisions in their cases; and that developing highly autonomous lethal weapon systems that outperform human soldiers would not create a morally objectionable responsibility gap in which no one could justly be held responsible for a weapon system's behavior. Boonin also argues that each of these conclusions can be successfully extended to a variety of other important cases. Written in a clear and accessible style, Artificial Intelligence, Ethics, and the State will be of interest to students and scholars in a variety of fields as well as to anyone concerned about the growing use of AI by governments around the world.