EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

XAI as Compliance and Design Strategy for Automated Grading Systems

Hasan Abu-Rasheed, Lezel Roddeck

Asia Pacific Journal of Educators and Education · 09.06.2026 · 41 (1) · 449-476

DOI: 10.21315/apjee2026.41.1.19 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Asia Pacific Journal of Educators and Education, online erschienen am 09.06.2026. Verfasst von Hasan Abu-Rasheed und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Ensuring explainability in Artificial Intelligence (AI) systems requires the careful integration of legal compliance and technical design. This paper proposes a compliance and design strategy for operationalising explainability obligations under the European Union’s Artificial Intelligence Act (AI Act), and the General Data Protection Regulation (GDPR), using automated grading systems (AGSs) in education as a case study. We argue that legal requirements and technical approaches should be coordinated from the earliest stages of system development to ensure trustworthy AI deployment. Our analysis combines legal analysis on the provisions of the AI Act’s relating to “high-risk” educational AI and the GDPR’s provisions on solely automated decision-making with a technical assessment of explainability methods including SHAP, LIME, feature sensitivity analysis, and interpretable models, and evaluate their appropriateness for educational contexts. An illustrative analysis of automated grading systems (AGSs) as a representative educational AI use case highlights how risks can emerge in automated grading and how targeted explainability approaches may mitigate these risks. Legal analysis reveals a duty to provide meaningful, fair and transparent explanations. A compliance and design strategy is proposed that integrates human-in-the-loop (HITL) mechanisms, collaborative design processes with educators, and stakeholder-centric explanation techniques. This interdisciplinary strategy demonstrates how technical methods and legal standards can work together to reduce risks, improve accountability, and strengthen trust in AGSs.

Zitieren

APA

Abu-Rasheed, H., & Roddeck, L. (2026). XAI as Compliance and Design Strategy for Automated Grading Systems. Asia Pacific Journal of Educators and Education, 41(1), 449-476. https://doi.org/10.21315/apjee2026.41.1.19

Vancouver

Abu-Rasheed H, Roddeck L. XAI as Compliance and Design Strategy for Automated Grading Systems. Asia Pacific Journal of Educators and Education. 2026;41(1):449-476. doi:10.21315/apjee2026.41.1.19

BibTeX
@article{AbuRasheed2026,
  author = {Abu-Rasheed, Hasan and Roddeck, Lezel},
  title = {XAI as Compliance and Design Strategy for Automated Grading Systems},
  journal = {Asia Pacific Journal of Educators and Education},
  year = {2026},
  volume = {41},
  number = {1},
  pages = {449--476},
  doi = {10.21315/apjee2026.41.1.19}
}
Kostenlos starten →