[加拿大]斯蒂芬·道恩斯
（卡塞尔曼，安大略 K0A 1M0，加拿大）
肖俊洪 译
【摘 要】人工智能是指能够完成传统上只有人类才能完成的工作的数字化工具。过去几年，人工智能在教育中的应用引发了广泛的讨论和试验。与此同时还出现了围绕人工智能教育用途的适切性和相关伦理的争议。有些人认为可以通过采用一套共同认可的原则（如可持续、问责和公平等）规范符合伦理的人工智能的教育用途。本文对这种观点提出了质疑。首先列举了九条人工智能伦理原则，并逐一进行了全面的思考和分析，指出其隐含的假设、矛盾之处和其他伦理方面的问题。文章认为这些原则所表达的大多数伦理考量，事实上是反映了某种政治主张和观点。关于人工智能教育用途的伦理，文章的结论是没有任何伦理原则能够获得各个利益相关方的一致认可。
【关键词】伦理；原则；人工智能；教育；公平
On Ethical AI Principles
Stephen Downes
(Casselman, Ontario K0A 1M0, Canada)
Abstract: Artificial Intelligence (AI) is a set of digital tools that can perform functions traditionally limited to human capability, for example, reviewing, summarizing, translating, and composing. In the last few years, there has been widespread discussion and experimentation for its application in education. Consequently, a contentious ethical debate has arisen around its appropriateness for educational functions and how such functionality may be ethically applied. Some argue that the ethical use of artificial intelligence in education can be defined through the use of a set of commonly held principles, such as sustainability, accountability, and fairness. This article challenges that presumption. A list of nine such principles is offered. Each principle is considered in detail and analyzed to identify underlying assumptions, points of conflict, and other ethical considerations. It is argued that much of what is offered as a set of ethical considerations reflects, in fact, a political argument and perspective. There is thus no set of ethical principles that can be regarded as a consensus opinion on the ethics of the use of artificial intelligence in education.
Keywords: ethics; principles; artificial intelligence (AI); education; fairness
下载： 
论人工智能伦理原则.pdf
