Human-Centric eXplainable AI in Education

Explainability is essential for building trust in educational AI: The paper emphasizes that as AI systems become more integrated into classrooms, users—both educators and students—need to understand how these systems work and make decisions. Without transparency, AI tools risk being viewed as "black boxes," which undermines trust and adoption in educational settings.

Different users have different needs for AI explanations: The paper identifies that educators, students, and administrators have diverse requirements for understanding AI systems. A one-size-fits-all approach to explainability won't work; instead, Human-Centric eXplainable AI (HCXAI) frameworks must be tailored to match the specific contexts and literacy levels of different stakeholders in education.

Implementation of explainable AI in education faces real complexity challenges: The paper highlights that creating understandable AI systems is difficult because educational AI models are inherently complex. Developers must balance the sophistication needed for effective learning tools with the simplicity required for users to actually comprehend how recommendations or assessments are being made.

Educators, developers, and policymakers must collaborate on ethical AI solutions: The paper calls for coordinated action across three groups to ensure AI in education is not only effective but also equitable and responsible. This includes establishing standards for explainability and ensuring diverse learners aren't disadvantaged by opaque AI systems.

Explainable AI can help achieve equitable educational outcomes: By prioritizing transparency and user understanding, HCXAI has the potential to support diverse learners and create more inclusive educational experiences, rather than perpetuating existing inequalities through unexplained algorithmic decisions.