Bỏ qua để đến nội dung

Course operations intervention and human grading governance

DomainsDOL EnglishProduct285 words1 min read
confirmedbyProduct Design

DEC-0013 - Course operations intervention and human grading governance

Phần tiêu đề “DEC-0013 - Course operations intervention and human grading governance”

Need explicit course-side role boundaries and review gates while keeping self-study fully automated.

CS intervention rule (course mode only):

  • CS intervenes every time learner is absent.
  • CS also intervenes when learner is late 3 times within a rolling 14-day window.
  • CS contacts learner (and parent when relevant).
  • CS first contact should happen within 48 hours after trigger.
  • CS supports makeup scheduling.
  • CS performs supportive check-ins to encourage attendance consistency.

AI Tutor role:

  • AI Tutor supports learners during learning.
  • AI Tutor can answer learner questions.
  • AI Tutor can summarize learning process.
  • AI Tutor can suggest suitable next learning actions.

Human grading governance for important tests:

  • The following test groups require teacher intervention before final scoring:
    • assignment,
    • level-check tests,
    • important in-course practice checkpoints,
    • final test.
  • Purpose: ensure result quality for high-impact evaluations.

Priority rule:

  • In course mode, teacher-reviewed outcomes are authoritative for important tests.
  • AI Tutor remains support layer, not final scorer for important course tests.

Visibility rule for learners:

  • Any information that does not require human intervention/review is shown immediately to learner.

Default visibility catalog:

  • Immediate (no human review needed):
    • attendance and schedule status updates,
    • regular progress counters/statistics,
    • AI Tutor responses and suggestions,
    • auto-scored non-critical practice results.
  • Review-gated (human intervention required):
    • assignment final scoring,
    • level-check test final scoring,
    • important in-course checkpoint final scoring,
    • final test final scoring,
    • any manually overridden score that changes official outcome.

Protect assessment quality in course mode while maintaining practical learner support through CS and AI.