Creating an Artificial Moral Agent

我建议对人工机器的道德能动性进行跨学科研究。在此，我将解决人工实体的道德责任问题，并致力于开发一种能够做出自己的道德判断的机器学习算法，这可以被认为是一个道德代理。在这一领域开展研究的重要性怎么强调也不过分。技术进步每周都会带来新的道德问题——无论是训练不足的机器学习算法基于种族或性别不公平地歧视某人，还是自动驾驶汽车导致死亡，没有人知道该归咎于谁。对伦理算法创建的研究已经开始，但直到最近才有人试图将这一创建过程正式化，并将哲学和伦理与人工智能的发展正确地联系起来。我的跨学科研究将进一步弥合这一差距，利用哲学和计算机科学的专业知识，为承认机器的道德责任而争论，开发一个可以在这样的机器中实施的道德框架，并创建一个将实现所选框架的算法。

I propose inter-disciplinary research focussed on the moral agency of artificial machines. Within this, I will tackle both the problem of moral responsibility in artificial entities, and work on developing a machine learning algorithm which is able to make its own ethical judgements, which could be considered a moral agent. The importance of developing research in this area cannot be overstated. New moral problems are presented weekly by technological advances - whether a poorly trained machine learning algorithm has unfairly discriminated against someone based on their race or gender, or a self-driving car has resulted in a fatality and nobody knows who to blame. Research into the creation of ethical algorithms has begun, however only recently have there been attempts to formalise this creation process, and to properly link philosophy and ethics with AI development. My inter-disciplinary research will further bridge this gap, drawing on expertise from both philosophy and computer science, to argue for the moral responsibility of machines to be recognised, develop an ethical framework that could be implemented in such a machine, and create an algorithm which will implement the chosen framework.