AI assists; accountable humans decide
AI can accelerate drafting, classification, quality checks, and analysis. It does not make an unreviewed draft authoritative or publish external claims without approval.
Permitted uses
Agents may cluster public keywords, draft briefs, propose practice items, summarize non-sensitive performance data, and recommend experiments. Outputs remain traceable to a task, model, prompt or policy version, and reviewer decision.
- Drafting from approved source material
- Structured quality and duplication checks
- Public search and campaign analysis
- Recommendation generation with cited evidence
- Synthetic examples that contain no student identity
Required safeguards
Student, counselor, sponsor, and institutional data defaults to approved protected processing. Public or cloud models receive only non-sensitive material. Educational answers, public claims, publishing, outreach, and ad spend require the applicable human approval gate.
- No PII in analytics or public-model prompts
- No autonomous ad spend or external outreach
- No fabricated sources or credentials
- Human review for anchored assessment content
- Model and policy evaluations before production changes
Primary references
Sources were checked on 2026-07-17. A link is evidence for the specific statement described, not an endorsement of ExamPrepPro.
Related standards and resources
See something we should correct?
Send the exact page or claim and supporting evidence to support@exampreppro.ph. Material corrections follow the public corrections policy.