Explanation as a Social Practice: Toward a Conceptual Framework for the Social Design of AI Systems

Abstract

The recent surge of interest in explainability in artificial intelligence (XAI) is propelled by not only technological advancements in machine learning, but also by regulatory initiatives to foster transparency in algorithmic decision making. In this article, we revise the current concept of explainability and identify three limitations: passive explainee, narrow view on the social process, and undifferentiated assessment of understanding. In order to overcome these limitations, we present explanation as a social practice in which explainer and explainee co-construct understanding on the microlevel. We view the co-construction on a microlevel as embedded into a macrolevel, yielding expectations concerning, e.g., social roles or partner models: Typically, the role of the explainer is to provide an explanation and to adapt it to the current level of understanding of the explainee; the explainee, in turn, is expected to provide cues that guide the explainer. Building on explanations being a social practice, we present a conceptual framework that aims to guide future research in XAI. The framework relies on the key concepts of monitoring and scaffolding to capture the development of interaction. We relate our conceptual framework and our new perspective on explaining to transparency and autonomy as objectives considered for XAI.

Trust, but Verify: Informed Consent, AI Technologies, and Public Health Emergencies

“…Taking this further and echoing [9] on trust, Israelsen and Ahmed, meanwhile, focus on trust-enhancing "algorithmic assurances" which echo traditional constructs like trustworthiness indicators in the trust literature (see Section 4.4) [11]. All of this comes together as positioning AI explainability as a coconstruction of understanding between explainer (the advanced AI-enabled technology) and explainee (the user) [12]. This ongoing negotiation around explanability echoes my own trust-based alternative to the dialogue around informed consent below (Section 4.4).…”

“…Much of the research above makes explicit a link between the motivation towards explainable or responsible AI with regulation and data subject rights [2,4,5,9,11,12]. With specific regard to big data, the Toronto Declaration puts the onus on data scientists and to some degree governance structures to protect individual rights [13].…”

“…[5])-share responsibility for the consent process. Even where studies have focused on stakeholder interactions and the co-construction of explainability [12], there is an evident need to consider the practicalities of the negotiation between parties to that process. For contact tracing, for example, who is responsible to the app user for the use and perhaps sharing of their data?…”

“…Leading on from studies which situate explainable AI within a behavioural context ( [6,11,12]), I take the more general discussion about the use and effects of advanced technologies into the context of planned behaviour (in Section 3) and extend discussions of trust [9,11] into the practical consideration of informed consent in a number of different domains. Starting with contact tracing and similar applications of AI technologies (Section 5), this discussion seeks to explore the confusion around consent in a practical context, evaluate the feasibility of transparency, and review responsible stakeholders for different scenarios.…”

“…Further, occasional (though infrequent [23]) re-affirmation of the relationship is advantageous. So, ongoing communication between the main actors is important in maintaining trust (see also [12]).…”