There are frequent calls for increasing ethical awareness within institutions that collect, aggregate, analyze and share large amounts of data. However there is little substantive guidance on how to put this into practice. This project, a collaboration with Accenture and the Atlantic Council, addresses how to build AI and data ethics capacity within organizations.
Topics, readings and speakers are decided upon by members of the group on an ongoing basis. Examples of topics include justice and fairness in machine learning, the form and extent of rights to information and technology access, the appropriate roles of institutions to prevent dissemination of misinformation, the responsible collection and sharing of data, AI research oversight models, ethical and philosophical issues in augmented and virtual reality, and the moral status of artificial intelligences. The group also aims to encourage and develop information ethics research projects and collaborations by its members. Students and faculty from any discipline are encouraged to join, though typically students will have taken at least one or more classes in philosophy before joining the group.
This report discusses the use of oversight committees and review boards help to identify and address issues in data ethics within institutional contexts. The report compares the relative merits of a review board system with other strategies for building ethical capacity, such as a predominantly compliance and training model. The project also details the components of an effective review board or committee based system and what is involved in building one.
This report discusses how to translate general AI and data ethics principles, which have been widely adopted by organizations, into action-guiding commitments. The report focuses on two common ethical commitments in AI and data use: a commitment to justice and a commitment to transparency. The report shows that the first step of fulfilling these commitments is specifying them into more substantive guidance and then operationalizing that guidance into concrete standards and practices.