TERI / Layer 1 — Foundations (Policy, Data & Systems) / 1.10
Checkpoint 1.10

1.10 — AI in Administrative Decision-MakingFrontier

What this is

Governance over the district's use of AI in administrative decisions that affect students and staff — behavior monitoring and student safety platforms (e.g., Gaggle, Bark, GoGuardian, Lightspeed), early-warning attendance / dropout prediction systems, hiring resume-screeners, scheduling and special-education AI, discipline-disparity flagging tools, and any algorithmic system that influences a high-stakes decision behind the scenes. Distinct from classroom AI use (1.3 / 1.4): this audit asks whether the district names, vets, monitors, and meaningfully reviews the AI inside its own administrative apparatus.

Why it matters

AI is increasingly used for high-stakes administrative decisions — flagging students for safety intervention, predicting dropout risk, screening teacher applicants, allocating resources. These systems are often built on biased training data. A district can have rigorous student-facing AI governance and still benefit from auditing the algorithms making consequential decisions about students behind the scenes.

Connects to

The Framework: Condition #8 (Strategic Tool Selection & Data Governance), Cognitive & Ethical Foundation — Ethical Reasoning Under Real Pressure. Links to 1.3 (AI Policy) which covers student-facing use, 1.5 (Data Governance) for the data flowing into these systems, and 1.6 (EdTech Vetting) for the procurement layer.

Maturity levels

Not Started
No district inventory of AI systems used in administrative decisions. Tools are adopted department-by-department without scrutiny of bias, accuracy, or appeal pathways. Affected students and families are typically unaware AI was involved.
Emerging
Some awareness that AI is in use administratively (behavior monitoring, attendance prediction). No formal inventory or governance — usage tracked informally if at all. No clear accountability for when AI gets a decision wrong.
Established
Formal inventory of AI systems used in administrative decisions, kept current. Each has a governance owner. Vetting includes bias, accuracy, and appeal-pathway analysis. Affected families notified that AI is in use. Periodic review of decisions where AI flagged students or staff. Human override required for high-stakes decisions.
Expanding
Continuous audit of AI administrative decisions disaggregated by demographic and outcome. Bias monitoring runs quarterly. Findings reported publicly to the board. AI systems retired when accuracy or fairness drops. The default for any new administrative AI tool is human-in-the-loop and transparent to families — not silent automation.

Go deeper with

Example resource
EDSAFE SAFE Benchmarks — administrative AI lens
Also consider