University AI Policy Tracker

Dossiers ouverts et fondes sur des preuves sur les politiques IA universitaires.

APIGitHub
Langue
ENZHFRPL

Analyse de politique

Profils d'analyse etayes, pas des conseils de politique

Dimensions derivees des dossiers publics de revendications et preuves.

Profiles788Schemauapt-policy-analysis-v1Review: Machine candidatePublic API/api/public/v1/analysis/index.json
788

profils d'analyse

6515

dimensions etayees

82/100

score moyen de couverture

57

langues sources conservees

What analysis means

Analysis profiles are derived metadata over public claim/evidence records. They summarize policy dimensions without replacing source evidence.

What coverage score means

Coverage score measures breadth of visible public evidence, not policy quality, safety, legality, strictness, or compliance.

How to cite analysis

Cite the analysis JSON with the related university record, basis claim IDs, source URLs, review state, and original evidence.

Summary

For citation context.

As of the current public release, University AI Policy Tracker has deterministic policy dimension analysis for 788 university records. 6515 records have source-backed evidence for this dimension; 228 records are classified as required. These analysis profiles are machine-candidate derived metadata and should be cited with their source claim IDs and public JSON.

Policy Coverage Score measures breadth of public, source-backed coverage. It is not a quality, strictness, legal adequacy, safety, or institutional compliance score.

Limite

Derived metadata with source links.

  • Dimensions summarize public tracker claims into consistent labels.
  • Original-language evidence remains canonical.
  • Confidence is separate from review state.
  • This tracker is not legal advice, not academic integrity advice, and not an official university statement unless a linked source is the university's own official page.
  • not_mentioned means no matching public tracker evidence is present in the current profile.
  • Open basis claim IDs and source URLs before reuse.

Dimensions

All dimensions are visible and contract-backed.

Policy presence

Whether the public record contains official AI policy or guidance sources.

703: Unclear85: Not Mentioned
Evidence-backed703Not mentioned85

AI disclosure

Whether public guidance addresses disclosure, acknowledgement, citation, or declaration of AI use.

270: Required266: Not Mentioned249: Recommended3: Unclear
Evidence-backed522Not mentioned266

Coursework

Whether public guidance addresses coursework, assignments, syllabi, papers, homework, or submitted work.

286: Restricted200: Required128: Conditionally Allowed62: Blocked
Evidence-backed753Not mentioned35

Exams

Whether public guidance addresses exams, tests, quizzes, or examination conditions.

355: Restricted206: Required64: Conditionally Allowed61: Blocked
Evidence-backed749Not mentioned39

Privacy and data entry

Whether public guidance addresses personal, confidential, sensitive, regulated, or student data entry into AI tools.

388: Restricted217: Not Mentioned80: Blocked43: Recommended
Evidence-backed571Not mentioned217

Academic integrity

Whether public guidance connects AI use to academic integrity, misconduct, dishonesty, plagiarism, or cheating rules.

228: Required219: Restricted115: Not Mentioned81: Conditionally Allowed
Evidence-backed673Not mentioned115

Approved tools

Whether public guidance identifies institutionally approved, licensed, procured, or enterprise AI tools.

202: Restricted194: Not Mentioned145: Allowed82: Required
Evidence-backed594Not mentioned194

Named AI services

Whether public guidance names AI services such as ChatGPT, Copilot, Claude, Gemini, Grammarly, or DeepSeek.

336: Restricted143: Not Mentioned93: Blocked59: Unclear
Evidence-backed645Not mentioned143

Teaching guidance

Whether public guidance addresses instructors, teaching, classroom policy, assessment design, or syllabus language.

628: Recommended160: Not Mentioned
Evidence-backed628Not mentioned160

Research guidance

Whether public guidance addresses research use, publication ethics, research data, grants, or human-subjects compliance.

333: Not Mentioned269: Recommended186: Restricted
Evidence-backed455Not mentioned333

Security and procurement

Whether public guidance addresses security review, procurement, vendor approval, risk assessment, authentication, SSO, or licensing.

566: Not Mentioned79: Restricted57: Required27: Recommended
Evidence-backed222Not mentioned566

Score de couverture

The score is a coverage breadth measure only.

Open coverage table

Coverage score components sum to 100 points across public evidence categories such as central guidance, academic integrity, disclosure, coursework, exams, privacy/data entry, approved tools, and teaching or research guidance.

The score should never be described as best policy, worst policy, most compliant, legally safe, or institutionally endorsed.

Seuils qualite

Publication gates keep analysis pages useful for search and AI answer engines without turning sparse data into thin pages.

Page-quality JSON

Current page-quality status: passes current quality gate. The report covers 5 public analysis pages and keeps review state separate from page publication readiness.

Server-rendered public content

Analysis conclusions, caveats, review state, evidence counts, and public JSON links must be visible in HTML.

Gateserver_rendered_public_content

Theme evidence threshold

Theme analysis pages are generated only when at least five records have source-backed evidence for the dimension.

Gatetheme_evidence_threshold

Review state boundary

Review state must be visible and must remain separate from confidence.

Gatereview_state_boundary

Original-language evidence

Source-language evidence remains canonical; localized display cannot replace it.

Gateoriginal_language_evidence

Coverage score caveat

Coverage scores must be described only as breadth of public source-backed coverage, not quality, compliance, legality, safety, or ranking.

Gatecoverage_score_caveat

No-advice boundary

Pages must state that tracker analysis is not legal advice, academic integrity advice, or an official university statement.

Gateno_advice_boundary

Versioned public JSON

Public analysis metadata must use /api/public/v1/... endpoints.

Gateversioned_public_json

Flux de revue

Quality checks do not publish canonical analysis. They route questionable records into review.

Open review workflow

Analysis profiles remain machine_candidate until reviewers confirm evidence binding, not-mentioned reasoning, source-language preservation, page quality gates, and no-advice boundaries.

Queueanalysis_profile_reviewPublic mutationNot allowedReview states6

Analyse thematique

Only themes with enough public evidence are linked from this index.

AI disclosure

Source-backed analysis of public university guidance on disclosing, acknowledging, citing, or declaring AI use.

Open analysis

Privacy and data entry

Source-backed analysis of public university guidance on personal, confidential, sensitive, regulated, or student data entry into AI tools.

Open analysis

Approved AI tools

Source-backed analysis of public university guidance naming approved, licensed, procured, or enterprise AI tools.

Open analysis

JSON public

Versioned read-only endpoints for downstream reuse.

Citation

Use the canonical page and public JSON together.

Suggested citation: University AI Policy Tracker. "University AI policy analysis profiles." Version v1. https://eduaipolicy.org/analysis

Tracker metadata uses CC-BY-4.0. Official source documents, source page text, PDFs, and other source materials retain their original rights and terms.