ABSTRACT

Evidence-based policymaking must urgently consider regulations addressing advances made in artificial intelligence (AI) and robotics, as well as issues of ownership, management and control. Many repetitive manufacturing tasks once requiring human labour have already been replaced by robots. The general public are facing associated risks to safety and privacy with the appearance of drones, driverless cars, robots of care, and human repair and enhancement. Major challenges for policymakers arise as machines acquire the ability to learn and become autonomous in their decision making. When independent of the humans that created them, their true ‘intelligence’ is tested in terms of their status as ‘moral beings’. Given their algorithmic decision making, questions arise about how ‘they’ could make ethical decisions about their actions and interactions with humans. This paper reviews current issues raised about the morality of autonomous learning machines and explores whether policy can be developed to address their potential to acquire a moral status. The authors argue that policies and regulations will fail if no account is taken of the ethics of robotics, either to ensure an ‘ethics by design’ or prevent them from becoming ‘autonomous moral agents’.

摘要

基于证据的决策必须紧急考虑针对人工智能（AI）和机器人技术的发展以及所有权，管理和控制问题的法规。曾经需要人工的许多重复性制造任务已经被机器人取代。无人机，无人驾驶汽车，护理机器人以及人类维修和保养的外观，给公众带来了与安全和隐私相关的风险。随着机器在决策过程中获得学习并变得自主的能力，决策者面临的主要挑战。当独立于创造它们的人时，他们的真正“智慧”就根据它们作为“道德生命”的地位而受到考验。根据他们的算法决策，关于“他们”如何做出关于他们的行为和与人类互动的道德决策的疑问。本文回顾了当前有关自主学习机器道德的问题，并探讨了是否可以制定政策来解决其获得道德地位的潜力。作者认为，如果不考虑机器人技术的伦理，要么确保“设计伦理”，要么防止它们成为“自治的道德主体”，政策和法规将失败。