https://ris.uni-paderborn.de 2000-01-01T00:00+00:00 1 daily https://ris.uni-paderborn.de/record/65063 Alpsancar, Suzana Stamboliev, Eugenia 2026 <jats:title>Abstract</jats:title> <jats:p> This chapter critically examines how social explainable AI (sXAI) can better support AI practitioners in ensuring fairness in AI-based decision-making. We argue for a fundamental shift: Fairness should be understood not as a technical property or an information problem, but as a matter of vulnerability—focusing on the real-world impacts of AI on individuals and groups, especially those most at risk. Hereby, we call for a shift in perspective: from fair AI to <jats:italic>tasking AI fairly</jats:italic> . To motivate our vulnerability approach, we review the “Dutch welfare fraud scandal” (system risk indication—SyRI) and current challenges in the field of fair AI/machine learning (ML). Vulnerability of a person or members of a definable group of persons is a complex relational notion, and not a technical property of a technical system. Accordingly, we suggest several nontechnical strategies that hold the promise to compensate for the insufficiency of purely technical approaches to fairness and other ethical issues in the practical use of AI-based systems. To discuss how sXAI, due to its interactive and adaptive social character, might better fulfill this role than current XAI techniques, we provide a toy scenario for how sXAI might support the virtuous AI practitioner in an ethical inquiry. Finally, we also address challenges and limits of our approach. </jats:p> https://ris.uni-paderborn.de/record/65063 eng Springer Nature Singapore info:eu-repo/semantics/altIdentifier/doi/10.1007/978-981-96-5290-7_29 info:eu-repo/semantics/altIdentifier/isbn/9789819652891 info:eu-repo/semantics/altIdentifier/isbn/9789819652907 info:eu-repo/semantics/openAccess Alpsancar S, Stamboliev E. Tasking AI Fairly. How to Empower AI Practitioners With sXAI? In: <i>Social Explainable AI</i>. Springer Nature Singapore; 2026:557-581. doi:<a href="https://doi.org/10.1007/978-981-96-5290-7_29">10.1007/978-981-96-5290-7_29</a> Tasking AI Fairly. How to Empower AI Practitioners With sXAI? info:eu-repo/semantics/bookPart doc-type:bookPart text http://purl.org/coar/resource_type/c_3248