Motivation
Traditional AI policies use a hierarchical inheritance model where permissions are grouped into rigid, cumulative levels. This creates friction when a learning activity requires specific permissions from different levels without adopting the entire tier. SCRIPT moves away from this “all or nothing” approach toward a compositional model. It treats AI permissions as discrete, modular components that can be mixed and matched to align with specific learning objectives.
Components
The framework categorizes AI usage into six components:
- Style: Modulating tone or stylistic voice.
- Correction: Fixing mechanical grammar, spelling, and syntax errors.
- Review: Receiving critical feedback or quality analysis on existing work.
- Ideation: Brainstorming, exploring topics, and generating initial ideas.
- Production: Generating primary artifacts like text, code, or images.
- Transparency: Mandatory disclosure of AI assistance and proper citation of content.
Implementation
Instructors select which components are authorized for a given task. This allows for granular control, such as permitting Production for code while strictly prohibiting it for prose. If a component is not explicitly authorized, its use is considered an infraction.
Example
Consider a master’s of education course called “Pedagogy of Digital Learning”:
| Task | Allowed | Disallowed | Reason |
|---|---|---|---|
| 1. Reflective Journal | Correction | Style, Review, Ideation, Production | This task requires an authentic writer’s voice to communicate personal thoughts and discovery. |
| 2. Survey of Tools | Review, Ideation, Correction | Style, Production | AI can assist in finding or critiquing tools, but the core research and decision making must be your own. |
| 3. Learning Activity | Style, Correction, Review, Ideation, Production | Production | You may use AI to create student materials, but you must independently justify the underlying pedagogy. |
| 4. Assessment Redesign | S, C, R, I, P | None | This task encourages co-designing with AI to evaluate its impact on assessment validity and learning. |
| 5. Community of Inquiry | None | S, C, R, I, P | This task evaluates social presence and human interaction, which AI usage would fundamentally undermine. |
Transparency (T) is mandatory for all AI assisted tasks.
AI Disclaimer
To develop this framework, AI was used for the Review and Ideation components. The original idea for composition vs inheritance is my own.