From What to How. An Overview of AI Ethics Tools, Methods and Research to Translate Principles into Practices

By Jessica Morley, Luciano Floridi, Libby Kinsey, Anat Elhalal. 2019

71 citations recorded by the source corpus.

The debate about the ethical implications of Artificial Intelligence dates from the 1960s. However, in recent years symbolic AI has been complemented and sometimes replaced by Neural Networks and Machine Learning techniques. This has vastly increased its potential utility and impact on society, with the consequence that the ethical debate has gone mainstream. Such debate has primarily focused on principles - the what of AI ethics - rather than on practices, the how. Awareness of the potential issues is increasing at a fast rate, but the AI community's ability to take action to mitigate the associated risks is still at its infancy. Therefore, our intention in presenting this research is to contribute to closing the gap between principles and practices by constructing a typology that may help practically-minded developers apply ethics at each stage of the pipeline, and to signal to researchers where further work is needed. The focus is exclusively on Machine Learning, but it is hoped that the results of this research may be easily applicable to other branches of AI. The article outlines the research method for creating this typology, the initial findings, and provides a summary of future research needs.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

The ethics of AI in health care: A mapping review. (2020, background): “Yet, research in these areas is currently happening in siloes, is overly focused on individual level impacts (Morley and Floridi, 2020a), or does not consider the fact that the ethical concerns may vary depending on the stage of the algorithm development pipeline (Morley et al., 2019).”
Principles alone cannot guarantee ethical AI (2019, unknown): “Many such methods exist 60 , including participatory design, reflective design, Values at Play and value-sensitive design 1 , but thus far they have largely been implemented and studied in academic contexts, which are more receptive to normative concerns than commercial settings 22,60,61 .”
Conversational AI: Social and Ethical Considerations (2019, background): “Engaging in use-case centred discourse can help to elicit social values that may then be used to inform the design of a speciﬁc agent’s response strategy, especially where variation of values across user groups is high (value pluralism) [33].”
AI Ethics - Too Principled to Fail? (2019, unknown): “provide an up-to-date overview of tools, methods, and research to translate AI Ethics principles into practice.(49) x Other value-specific methods, such as ‘privacy by design’ or ‘security by design’ are more successful in this regard, in part because they address technically tractable normative con”