How do we show OCR we measured our school AI for discrimination?

You show it with documented, reproducible measurement of the tool's outcomes across protected groups, plus the mitigation steps you took where disparity appeared — evidence that you proactively assessed discriminatory impact rather than assumed neutrality.…

register 09 · Compliance pins· OCR AI Guidance
01 ·

Answer.

Education / student-data rule · OCR AI Guidance.

You show it with documented, reproducible measurement of the tool's outcomes across protected groups, plus the mitigation steps you took where disparity appeared — evidence that you proactively assessed discriminatory impact rather than assumed neutrality. Comparable, tamper-evident records are persuasive. Planisphere produces that cohort-comparison evidence as a sha-pinned record, run without ingesting student PII where designed; it evidences the measurement, OCR makes any compliance finding.

02 ·

The mark behind the answer.

OCR AI Guidance protects students — their records, their civil rights, their access. When an AI sits in an eligibility, …

ED Office for Civil Rights · discriminatory-AI nondiscrimination resource.

→ Full reference for OCR AI Guidance

Try OCR AI Guidance with a free test key.

Start with the free sandbox. Planisphere measures model behaviour and emits a reproducible, sha-pinned record — it does not certify, file, or give legal advice.

API ·

When OCR AI Guidance asks for proof, hand over records — not assurances.

record the duty · seal the receipt · verify offline

Planisphere records each duty event — an output marked, a disclosure shown, a review made — and seals it into a receipt that verifies offline against our published keys. You send hashes, never content. A record is evidence a third party can check; it is not a certification and not a legal determination.

See how a record is checked