Rethinking Higher Education: From Fixed Curricula to Learnity Graphs

Synthesis: Szekely, Gal-Ezer & Harel (2026) argue that AI-mediated knowledge access warrants rethinking fixed higher-education curricula, proposing "learnity graphs" — structured representations of learning as interconnected units of knowledge, skills, experience, and artifacts — as a lifelong-learning framework that integrates academic, professional, and personal learning.

Key Findings

1. The value proposition of universities persists. The authors emphasize that universities remain essential for foundational knowledge, theoretical depth, and conceptual grounding; the goal is not to replace academic study but to extend it.
2. AI mediation changes the calculus. In an era where knowledge is increasingly AI-mediated and accessible, the fixed-sequential-curriculum model is open to reconsideration, shifting emphasis toward Creativity, interdisciplinary integration, hands-on experience, and long-term development.
3. Learnity graphs as the core construct. The paper introduces learnity graphs, a structured representation of learning as interconnected units of knowledge, skills, experience, and actual artifacts, together with a method for presenting and leveraging the representation.
4. Lifelong integration across domains. The framework integrates academic, professional, and personal learning, positioning learning as a continuous, graph-structured activity rather than a bounded curricular sequence.

What this means for practice

• Curriculum designers. Represent programs as learnity graphs — nodes for knowledge, skills, experience, and artifacts, joined by typed edges for prerequisite, interdisciplinary, specialization, and compositional relations — instead of fixed unit sequences.
• Curriculum designers. Publish the representational layer before scaling: a shared language defining learnities, their relationships, development levels, contexts, and evidence types is the authors' stated precondition for cross-institutional use.
• Learning designers. Build environments in which learners navigate and maintain their own graph, since the framework locates value in the unique structure of an individual trajectory rather than in possession of knowledge.
• Learning designers. Keep the structure flexible enough to guide rather than constrain, because the authors warn that overly rigid standardization reproduces the limitations of fixed curricula.
• Administrators. Budget for the governance layer alongside the technology — knowledge representation, evaluation mechanisms, scalability, and cross-institutional recognition of learning — and for the lifelong advisory structures that make a learner's graph portable.

Limitations

• This is a conceptual framework paper: the authors state that implementation is left for future work, so no learning-outcome claim is tested or testable from this text.
• The only demonstrated artifact is a prototype system (learnity-graph.vercel.app), which shows representation, personalization, development tracking, and recommendation — there is no sample, comparison condition, or measured effect.
• The conditions under which the framework could be evaluated are themselves unresolved: knowledge representation, evaluation mechanisms, technological scalability, and governance for cross-institutional recognition.
• No criteria are given for granularity — when a learnity becomes a distinct node in the graph — which the authors acknowledge is required to avoid graph inflation and preserve interpretability.

Citation

Szekely, S., Gal-Ezer, J., & Harel, D. (2026). Rethinking Higher Education: From Fixed Curricula to Learnity Graphs. (cs.CY).