A Genuine Tension
Bringing AI into education creates a genuine tension between two equally reasonable positions. Administrators look at the technology and see real potential in automated early alerts, intelligent scheduling, and enrolment models that can be acted on weeks in advance. Faculty look at the same technology and see real risk in algorithmic decisions made about students they teach and automated communications written without the nuance a human in the room would bring. Both reactions are valid, and pretending the tension does not exist is the surest way to lose both audiences.
Progressive autonomy resolves this. AI capabilities exist on a spectrum from observe to suggest to act, and the humans governing the institution control where each capability sits. When first deployed, RIBA operates in observe mode. It watches, learns, and waits.
Observe, Suggest, Graduate
After observation, RIBA suggests. An advisor sees: 'Three students show engagement patterns consistent with withdrawal risk.' The advisor reviews the evidence and acts on their own judgment. If they find suggestions valuable, they grant permission to advance.
- The registrar might grant autonomy over scheduling conflicts
- Faculty might keep suggest-only for anything touching assessment
- The dean might allow autonomous report generation but require approval before external communications
- Every suggestion, action, and override is recorded in an immutable ledger
The governance is role-specific. Every role defines its own comfort level with AI assistance.
The Mentor Can't Be Automated
The audit trail becomes particularly important where decisions affect student outcomes. Every suggestion, every action, every approval or override is recorded in an immutable ledger. When an accreditation body asks about AI use in advising, the institution produces a complete, verifiable record.
Education will never fully automate the student experience. The mentor who notices a student struggling from the look on their face. The advisor who knows a direct conversation is needed. Progressive autonomy handles pattern recognition at scale while keeping consequential decisions with the people who chose education because they care about students. UNESCO's guidance on AI in education is explicit that human-led, AI-augmented learning environments are the only acceptable design pattern at consequence-bearing decision points. To run this on your institution's stack, get early access.
