| The rapid development of artificial intelligence technology has shaped humanmachine joint actions in a wide range of fields.Due to the complexity and specificity of artificial intelligence technology,the traditional responsibility attribution system faces great challenges in this context.For example,in human-machine joint actions,there are problems such as unclear responsibility subjects and difficult attribution.Therefore,"who is responsible for the decision-making of the human-machine joint system" has become the focus of debate.For the philosophical community,when we try to explore and clarify responsibility issues,we tend to focus more on moral responsibility.Based on the issue of moral responsibility attribution in human-machine joint action,the research goal of this paper is to sort out and analyze the general views on the allocation of responsibilities for human-machine joint action in the ethical guidelines for artificial intelligence established in today’s world,to explore whether this general view is reasonable,and to try to provide reference principles and ideas for clarifying the responsibility subjects and dividing the domain of human-machine responsibility within the joint system.The significance of this paper lies in that,from a theoretical level,it summarizes and sorts out the overall view of responsibility attribution for human-machine joint action in today’s world;from a practical level,it proposes corresponding suggestions for the irrationality in the overall view,which can provide improving direction for establishing and perfecting ethical guidelines for artificial intelligence in the future.The improvement of guidelines will also provide ideas and basis for solving the problem of human-machine responsibility allocation.To achieve the research goal,this paper uses 85 artificial intelligence ethical guidelines issued from 2016 to February 2022 as research materials and adopts research methods such as keyword method and literature research method.The specific content of the research is divided into three parts:First,keyword search is conducted on the ethical guidelines,and analyzes each in combination with the context to explore the responsibility attribution it emphasized.Construct a hierarchical classification diagram of responsible subjects to comprehensively evaluate the responsible parties concerned by the 85 guidelines.Second,in view of the fact that artificial intelligence is rarely considered as a responsible party in ethical guidelines,this paper refutes from the necessity and possibility of attributing moral responsibility to artificial intelligence.From the perspective of necessity,it is explained that only attributing to humans will cause influence such as "responsibility gap",over-expectation of human subjects and inhibition of developers’ innovation ability;from the perspective of possibility,it refutes that having intentionality and free will equivalent to humans is a prerequisite for artificial intelligence to be morally responsible,and proposes ways to hold artificial intelligence accountable on a moral level from a practical perspective.Thirdly,based on the above arguments,this paper proposes that when formulating and improving the responsibility context of artificial intelligence ethical guidelines in the future,in terms of overall responsibility allocation issues,considering intelligent machines bear moral responsibility is possible and classified manner should be formulated.In solving the difficulty of defining responsible subjects,transparent communication and reporting mechanisms,traceability mechanisms and audit mechanisms should be established.When defining the respective responsibility categories of human-machine subjects,consideration should follow the principles of improving the decomposability of responsibility,following realistic experience and following the principles of intelligence level of intelligent machine and social division of labor.This paper has two innovations:First,it summarizes the responsible subjects emphasized by the 85 artificial intelligence ethical guidelines and their focus,and explores their irrational aspects.Second,it proposes measures to clarify the responsible subjects in human-machine joint actions and ideas for defining the scope of humanmachine responsibility within the system to improve the relevant content of responsibility in artificial intelligence ethical guidelines. |
