XAI-Ed 2027: Closing the Interpretation-Action Loop
Actionable Explainable AI to Enhance Learning and Teaching
The 4th International Workshop on Explainable AI in Education, at LAK27
Date: March 2027, full-day pre-conference workshop (exact day TBA)
Location: Recife, Brazil
Part of LAK27, the 17th International Learning Analytics & Knowledge Conference
Registration to the main conference is required to attend
The day at a glance
A full-day workshop that alternates paper sessions with hands-on design activities, all working from a shared scenario pack.
09:20
Invited talk
10:00
Paper session 1
11:15
The explanation clinic
13:30
Papers & demo carousel
14:15
Cross-stakeholder panel
15:30
Explanation-to-action chain
16:30
Research agenda
What counts as evidence that an explanation worked?
Students, teachers, caregivers, and educational leaders approach the same inference with different goals, prior knowledge, and decision rights. The panel debates what evidence shows that an explanation helped each of them learn, decide, or act.
Panelists to be announced soon
Read more
Why attend XAI-Ed 2027?
The 2025 revision of the definition of learning analytics places the interpretation and communication of data about learners alongside its collection and analysis, and names theoretically relevant and actionable insights that enhance learning and teaching as the purpose of the field. Explainability is the mechanism through which that interpretation and communication happen. Yet evidence about what explanations accomplish in education is still thin. This edition of XAI-Ed asks two questions: how explanations affect learning processes and outcomes, and how they become actionable for students, teachers, parents and caregivers, and educational leaders. Equity, fairness, and transparency run through both.
Goals and Objectives
We treat an explanation as a pedagogical artefact whose value lies in what changes in learning and in decision-making when it is present, absent, or differently formulated. The workshop aims to:
- Consolidate methods for studying how explanations affect learning
- Develop measures of actionability that separate understanding an inference from acting on it
- Articulate stakeholder-differentiated explanation design
- Establish explanatory equity and contestability as design and evaluation criteria for explainable learning analytics (XLA)
Expected Outcomes
A full day of paper and demonstration sessions, two structured design activities, and a cross-stakeholder panel, all working from a shared scenario pack, to produce:
- Explanation design cards for different stakeholders
- A map of where the explanation-to-action chain breaks, with each failure point turned into a research question
- A shared framework for actionable explainable learning analytics
- A community research agenda, continuing the XAI-Ed series
Interested in Contributing?
Submit your research or system demonstration to XAI-Ed 2027 and help close the loop between explanation and action in learning analytics.
Call for Papers