ABSTRACT
Transparency, explainability, and interpretability have been identified among the core principles of developing ethical AI-powered technologies. Despite the enormous relevance of these issues, most research around AI transparency, explainability, and interpretability remains constrained to technical disciplines. In this context, the explainable AI (xAI) field seems to dominate the conversation. The search for methods to make AI transparent to human understanding and tackle the reasoning behind the model’s decision are common concerns of researchers. Inspired by the need to bring more social sciences insights into this conversation, this chapter delves into how socio-constructivists’ approaches to technology and public engagement with science studies can contribute to broadening the understanding of AI transparency, specifically for the governance of AI in biomedical settings. The chapter includes recommendations and a box of questions created by the author to contribute to adopting this approach.
