Ethical reasoning in artificial intelligence: A cybersecurity perspective
本文提出一套伦理原则、推理方法和培训能力，帮助网络安全专业人员在设计人工智能系统时融入伦理考量，确保技术选择兼顾正负伦理影响。
The paper provides solutions to the main challenge of using cybersecurity and artificial intelligence with ethics in mind: ensuring that ethics is present and justified in the design process. We propose ethical principles, a reasoning method, and training competencies to achieve this goal. These principles are autonomy, beneficence, non-maleficence, trust, justice, and transparency. The competencies include defining the context, identifying ethical priorities, setting minimum ethical thresholds, and analyzing tradeoffs of prioritized ethical principles. The focus is on reasoning, which is justifying technical choices while considering their negative and positive ethical implications in context, from personal to social. Ethical reasoning also involves “ethical action” competencies (practical knowledge sets) that cybersecurity professionals and students can learn to incorporate ethical principles into their decision-making process.