Campus as an AI Sandbox: Open-Source Smart Building Infrastructure for Agentic AI Education
An open-source, five-layer smart-building platform turns campus infrastructure such as lighting, HVAC, elevators and energy systems into a student-modifiable environment for project-based agentic AI learning.
Degree of pedagogical or technical divergence from conventional educational technology patterns.
Development cycle: early signal → experiment → classroom pilot → scaling → established.
Rigour of published findings: anecdotal → single case → emerging cohort → moderate RCT → strong replication.
Core Educational Problem Addressed
Abstract STEM exercises disconnected from tangible real-world systems and immediate ecological consequences.
Target Learner Groups
Empirical Evidence Summary
Official EDUCAUSE session materials describe the five-layer platform and student-designed/deployed agentic AI applications across campus buildings. The public session page does not report comparative learning-outcome metrics.
Known Limitations, Constraints & Failure Modes
Demands close collaboration with facility management, strict student data privacy partitioning, and robust API maintenance.
Pedagogical Mechanism Extraction
This innovation decomposes standard instruction by substituting passive receptive media with active student verification cycles. The underlying dynamic shifts cognitive effort from rote recall toward epistemic evaluation of system outputs.
Technological Architecture
Operates via localized edge models and sensor event telemetry. Decoupled from proprietary cloud monopolies to minimize data leakage and latency spikes during in-class student interaction.
Relevance to Swiss Educational Context
High alignment with Swiss MINT initiatives and Lehrplan 21 interdisciplinary "Natur und Technik" (Sek I/II), especially for cantonal gymnasium living labs.
Exploratory Application Hypothesis
Instrument a pilot classroom cluster in a Bernese vocational school with open IoT telemetry; students examine heating efficiency trade-offs during winter quarters.
Network Lineage & Related Signals
Signals connected by pedagogical analogy, shared technical architecture, or cluster membership.
Mission-Based AI Skills Studio (Artifact-Centric)
Illustrative relationship for demonstrator UI — Shared pedagogical premise: replacing passive multiple-choice compliance with authentic physical or digital artifacts generated in living environments.
Practitioner & Expert Peer Reviews
Evaluations by PHBern researchers, school leaders, and canton educators assessing actual classroom feasibility and evidence rigor.
No expert evaluations filed yet for this signal. Qualified educators may contribute above.