Knowledge Observatory

Explore the constellation.

A living map of AI fluency, metacognition, assessment, coaching, learning systems, and workforce architecture. Each point is a claim, risk, product response, or unanswered question about whether AI is building capability or dependency.

Signals worth building from.

These entries are the live, approved surface. Weekly automation should produce drafts for review before anything is promoted here.

2026-05-01

Research signal

Knowledge graphs are becoming the trust layer for AI learning systems.

Recent education research is moving beyond chat interfaces toward hybrid systems that combine LLMs with structured knowledge graphs, retrieval, and explainability mechanisms.

Philomathia should treat the graph as more than a visual metaphor: it can become the model that links assessment evidence, coaching context, and workforce capability.

Graph RAG for learningExplainable tutoringBehavioural assessment

2026-05-01

Research signal

The AI productivity story is turning into a capability story.

Future-of-work research is increasingly focused on uneven benefits, collaboration patterns, and the redesign of work systems rather than raw task acceleration.

The strongest Philomathia position is not "AI makes people faster." It is "AI changes which human capabilities matter, and how organisations should cultivate them."

AI changes the work systemLive capability frameworksHuman-AI collaboration

2026-05-01

Research signal

Assessment needs to move from recall to observed reasoning.

If AI use can mask weak judgment, assessments that only check conceptual knowledge will become less useful exactly when organisations need better evidence.

Dojo should foreground live work traces: prompt choices, verification behaviour, adaptation, token efficiency, and recovery from misleading model output.

Behavioural assessmentToken efficiencyAssessment validity