Student AI Learning Compact
SALC Governance Framework
Version 1.0 | March 2026
Seven domains of principled standards for AI interaction with K-8 students. Convened by the Fulcra Institute.
These standards are written as a covenant — a shared commitment among districts, platforms, schools, and families to the children in K-8 learning environments. They are not a compliance checklist. They are the floor below which no SALC signatory may fall, and the foundation upon which every implementation must be built.
The Seven Domains
Legal and Compliance
Every SALC-compliant platform builds child safety law into its architecture, not its handbook.
Content Guardrails
The AI operates within defined topical boundaries appropriate to the child, with redirection that never shames.
Interaction Boundaries
All AI systems stay within their defined role; dialogic AI asks rather than tells.
Safety and Escalation Protocols
No child who discloses distress is left alone with an AI.
Session Structure Rules
Every session is purposeful, time-bounded, and designed for the developmental reality of the student in it.
Transparency Requirements
Students, teachers, and families know what the system is, what it captures, and what it does with it.
Equity and Access Standards
The framework is only as strong as its weakest implementation.
These seven domains represent the Fulcra Institute's founding standards for SALC. They are reviewed annually by the coalition governance body, with revisions published as versioned updates. Version 1.0 of these standards takes effect upon the launch of the SALC coalition at teachingwithai.org.
Framework Extensions
Extensions to the SALC address specific operational contexts not fully covered by the seven core domains. They are published under the same license and carry the same institutional weight.
AI Response & Reporting Protocol
Standards for how AI dialogue systems should respond to unexpected, inappropriate, or distressing student input — and how such events should be communicated to teachers. Defines three governing principles and six operational standards including tiered response classification, pre-defined safety responses, and blocklist governance.
Read Extension