Body
University AI Tiers
Below represents the 4 AI Tiers at the University based on the capabilities provided by the AI, the Human Approval Level needed, and the Governance Path required for usage.
|
Tier
|
Capabilities
|
Human Approval Level |
Governance Path
|
|
1: Informational AI
|
- Generates content
- Answers questions
- No system action authority
|
Level 0-1
Read Only, Recommendations
|
Light Governance
|
|
2: Assisted Enterprise AI
|
- Embedded in enterprise platforms
- Human-in-the-loop
- Recommendations and copilots
- Limited automation
|
Level 1-2
Recommendations,
Human Approval Required
|
Information Access Governance
|
|
3: Agentic Operational AI
|
- Executes workflows
- Uses tools/connectors
- Interacts with systems
- Can perform actions autonomously or semi-autonomously
- May access local files, browser sessions, APIs, code, cloud systems, computers, or development environments
|
Level 2-3
Human Approval Required, Limited Autonomous Actions
|
Action Governance
|
|
4: Autonomous Institutional AI
|
- Coordinates work across multiple university systems or data domains (orchestration)
- Multi-agent coordination
- Performs institution-level workflows with limited to no human intervention
- Can make or trigger operational decisions
- High impact to university operations, services, or data
|
Level 4
Autonomous Actions
|
Action Governance + Executive Review
|
University AI tools may move between tiers over time as new capabilities are enabled, particularly around workflow execution, connectors, agent orchestration, system access, and autonomous action-taking.
Governance decisions will consider enabled functionality and operational authority, not only vendor or product name.