From Ethical Principles to the Ethically-Informed Engineering of Autonomous Systems

The many sets of high-level ethical principles for AI and autonomous systems indicate a widespread societal concern about the ethical implications of these technologies. But, despite some notable examples, operationally-specific practical guidance on how to address this concern, during the design and engineering of the systems, is not yet at a mature stage. 

On 22 January 2021, the AAIP held an Ethics Workshop to contribute to the maturing of, and critical reflection upon, that practical guidance.

Participants from three communities were invited: technical; ethics; and regulatory policy. The format was a discursive, multi-disciplinary knowledge exchange on four questions:

1. What are the best ways to translate ethical principles into ethically-informed engineering practice?
2. Should the same core ethical principles be included in the decision-making process of every development team?
3. How should engineers approach ethics for, and under, uncertainty? 
4. Is an ‘ethical kitemark’ a good idea for systems developed by teams who have done their best to consider ethical impact?

This report details the answers proposed in response to the four questions. It also gives details on further themes that emerged throughout the discussions and outlines the next steps for AAIP research activity.