Design Principles for Explainable AI in Finance: A Multi- Stakeholder Framework

Christoph Kreiterling
Additional contact information
Christoph Kreiterling: Trier University of Applied Sciences

Working Papers from HAL

Abstract: Electronic marketplaces increasingly deploy AI for critical decisions, onboarding, pricing, approvals, and fraud detection. Effective explainable AI (XAI) must satisfy diverse stakeholders while maintaining technical accuracy. This research presents qualitative methods analyzing 24 semi-structured interviews across five specialized groups: retail users, advisors, developers, risk officers, and regulators from nine European countries. The investigation examined lending and buy-now-pay-later platforms through two scenarios - credit-limit changes and fraud-flag reviews - using validated XAI evaluation techniques. Three patterns emerged. First, process-oriented explanations that explain "why this case" enhance fairness perceptions and clarify actions. Second, progressive disclosure frameworks - beginning with concise text, then offering detailed visualizations - optimize comprehension without overload. Third, raw confidence metrics create uncertainty, whereas counterfactual examples demonstrate decision-altering factors effectively. These patterns, derived through systematic thematic analysis, challenge conventional approaches that prioritize algorithmic transparency over stakeholder comprehension. The study establishes six design principles for interpretable AI and proposes a multi-stakeholder evaluation framework connecting explanation tools to technical accuracy and human-centered outcomes. This responsible AI approach provides specialized meaning through role-appropriate explanations. The governance toolkit comprises operational performance indicators and audit-compliant documentation protocols, offering practical implementation methods for platform operators and regulators navigating XAI deployment in financial marketplaces.

Keywords: explainable AI in finance; electronic marketplaces; role-aware evaluation framework; counterfactual explanations; credit-limit decisions; fraud-flag reviews (search for similar items in EconPapers)
Date: 2025-09-01
Note: View the original document on HAL open archive server: https://hal.science/hal-05233140v1
References: Add references at CitEc
Citations:

Downloads: (external link)
https://hal.science/hal-05233140v1/document (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:hal:wpaper:hal-05233140

DOI: 10.13140/RG.2.2.13686.46408

Access Statistics for this paper

More papers in Working Papers from HAL
Bibliographic data for series maintained by CCSD ().