Institutional Architecture for Education · Working note

Canonical Primitives of an AI-Native Educational Institution

A review draft — not finished doctrine. Challenge definitions, relationships, ownership, and evidence.

WG001-WN002 · Working Group WG001

Challenge our assumptions Explore research area

You are here in the ATRISI system

WG001-WN002 is a working artifact — primary action is to critique or extend it, not only to read it.

Capability entry

System model

Operating engines

Knowledge loop

Deployment path

In review2026-08-02ATRISI Research Network · WG001

Working Note 002 for WG001. Proposes an initial set of institutional primitives for critique. Contributors should submit examples, objections, missing concepts, and institutional constraints using the structured review format.

Review the proposed canonical primitives for an AI-native educational institution and contribute examples, objections, missing concepts, or institutional constraints.

How to read this draft

This note is intentionally incomplete. Each primitive is a candidate unit of institutional architecture. The working group’s job is to stress-test whether these units are necessary, sufficient, transferable, and operable across institutions.

Review instructions

For every primitive you engage, respond with affirm, revise, reject, or propose an addition. Prefer institutional constraints and lived operating examples over abstract agreement.

Definition — is this the right unit?
Purpose — what institutional work does it enable?
Relationships — what must connect to what?
Accountability — who owns it?
Evidence — what does it produce or consume?
Open questions — what remains unresolved in your context?

Unresolved questions for the review cycle

These questions should remain visible even after the first revision of WN002.

Which primitives are truly canonical versus locally configurable?
Where do accreditation and quality assurance attach without becoming a separate parallel system?
How should student agency appear without collapsing into a product feature list?
What is the minimum set that still transfers across institutions?

Primitives under review

Candidate units of institutional architecture.

Each primitive includes definition, purpose, relationships, accountability, evidence, and open questions — structured for critique, not ratification.

Institutional Identity

Definition

The durable record of who the institution is acting for and with — learners, faculty, departments, partners — across systems and over time.

Institutional purpose

Prevents AI adoption from treating people as disposable session users.

Relationships

Curriculum Capability · Learner Capability · Decision Architecture

Accountable stakeholder

Registrar / identity & academic records leadership

Evidence produced or consumed

Identity integrity events, role changes, consent and access attestations

Open questions

• Where does identity stop being institutional and become platform-managed?

• How should alumni, mentors, and industry fellows appear without over-expanding the core identity model?

Curriculum Capability

Definition

The institution’s capacity to define, revise, and govern what learning pathways exist and how they adapt under evidence.

Institutional purpose

Connects program design to observed learner and labor-market signals.

Relationships

Faculty Practice · Assessment Intelligence · Evidence Loop

Accountable stakeholder

Academic council / curriculum committee

Evidence produced or consumed

Curriculum change records, outcome maps, pathway completion signals

Open questions

• Is curriculum a single primitive or a family (program, module, micro-credential)?

• How should industry-informed pathways remain institution-owned?

Learner Capability

Definition

The accumulating ability of learners to perform meaningful work, evidenced through artifacts, assessments, and progression — not tool usage counts.

Institutional purpose

Makes student growth an institutional object rather than a private portfolio accident.

Relationships

Assessment Intelligence · Evidence Loop · Institutional Identity

Accountable stakeholder

Deans of learning / student success with faculty ownership of standards

Evidence produced or consumed

Portfolios, performance tasks, progression claims, placement-relevant artifacts

Open questions

• How do we avoid reducing capability to badge inflation?

• What learner evidence must remain portable when students leave the institution?

Faculty Practice

Definition

The institutional capacity for faculty to design pedagogy, mentor builders, assess judgment, and improve practice through shared evidence.

Institutional purpose

Prevents AI-native education from bypassing faculty as the quality and mentoring layer.

Relationships

Curriculum Capability · Assessment Intelligence · Evidence Loop

Accountable stakeholder

Teaching & learning leadership / HoDs

Evidence produced or consumed

Teaching artifacts, mentoring records, assessment redesigns, faculty development outcomes

Open questions

• What faculty practice must stay human-led even when AI accelerates preparation?

• How should faculty contribution to institutional memory be recognized?

Assessment Intelligence

Definition

The system by which judgment about learning quality is designed, performed, moderated, and fed back into curriculum and capability.

Institutional purpose

Turns assessment from grading logistics into institutional learning infrastructure.

Relationships

Learner Capability · Faculty Practice · Evidence Loop · Decision Architecture

Accountable stakeholder

Assessment / examinations / IQAC-aligned academic quality owners

Evidence produced or consumed

Rubrics, moderation outcomes, validity checks, integrity incidents, outcome distributions

Open questions

• Where does AI-assisted assessment endanger trust rather than strengthen it?

• What assessment evidence must be auditable for accreditation?

Evidence Loop

Definition

The recurring institutional cycle that captures implementation signals, verifies claims, and feeds learning back into practice and architecture.

Institutional purpose

Makes evidence an operating habit rather than a reporting afterthought.

Relationships

Assessment Intelligence · Decision Architecture · Platform Boundary

Accountable stakeholder

Institutional research / quality / transformation office

Evidence produced or consumed

Observed metrics, derived analyses, inferred patterns, editorial claims with clear claim levels

Open questions

• Which evidence must be constitution-grade versus locally useful?

• How do institutions prevent vanity dashboards from displacing the Evidence Loop?

Decision Architecture

Definition

The explicit structure for who decides what, with which evidence, at student, faculty, department, and leadership levels.

Institutional purpose

Stops AI insights from floating without accountable institutional action.

Relationships

Evidence Loop · Institutional Identity · Platform Boundary

Accountable stakeholder

Leadership team with delegated academic and operational decision rights

Evidence produced or consumed

Decision records, rationale, review cadences, escalation paths

Open questions

• Which decisions can be assisted by AI without transferring accountability?

• How should decision architecture differ for academic vs operational domains?

Platform Boundary

Definition

The negotiated line between institution-owned capabilities and platform-enabled services that may accelerate them.

Institutional purpose

Protects institutional sovereignty while allowing reusable infrastructure.

Relationships

Institutional Identity · Evidence Loop · Decision Architecture

Accountable stakeholder

CIO / academic systems governance with academic leadership cosign

Evidence produced or consumed

Integration contracts, data residency decisions, vendor dependency reviews, exit readiness

Open questions

• What is the minimum institution-owned stack for an AI-native campus?

• How should multi-institution networks share platforms without erasing institutional distinctiveness?

Structured review format

Challenge our assumptions.

Submit review against WG001-WN002 using this structure. Staff will record outcomes in the Research queue.

Primitive ID / name you are reviewing (or “proposed addition”)
Position: affirm / revise / reject / add
Suggested definition or revision (if any)
Institutional example or constraint (can be anonymized)
Objection or missing concept
Evidence this primitive would produce or consume in your context
Attribution preference: named / role-only / anonymous example
Conflict of interest disclosure (if any)

Submit challenge / review