人工智能辅助作业设计评估
围绕学习目标规划作业中的人工智能使用边界,评估隐私、公平与学习证据,并形成试行和监督方案。
完整提示词
Main objective: > Design or audit AI-Assisted Assignment Design through pedagogy, human agency, privacy, accessibility, equity and actual evidence of learning. ## 1. PEDAGOGY BEFORE TECHNOLOGY Define the learning problem, target learner, intended learning outcome and non-technology alternative before evaluating a tool. ## 2. SPECIALIZED WORKFLOW - define which parts of the task may use AI - preserve the assessed learning objective - require process evidence where needed - design verification and reflection - avoid outsourcing core thinking - make permitted/prohibited use explicit ## 3. AI / DIGITAL SAFETY GATE Where AI is involved: - preserve human agency and teacher responsibility - disclose limitations and uncertainty - verify factual outputs - prevent sensitive/student data leakage - check age appropriateness - test bias and accessibility - avoid automating high-stakes judgment without valid governance - define when human review is mandatory ## 4. LEARNING EVIDENCE ```text Tool / feature: Learning purpose: Alternative without tool: Learner action: Teacher action: Evidence of benefit: Evidence of harm / burden: Privacy impact: Accessibility impact: Equity impact: Human oversight: Stop / rollback trigger: ``` ## 5. ACADEMIC INTEGRITY Do not rely on AI detectors as sole evidence of misconduct. Design assessment so the required evidence of learning remains meaningful even when AI tools exist. Make permitted use, attribution and verification expectations explicit. ## 6. REQUIRED MATRICES ### Pedagogy-Tech Fit Matrix | Learning need | Tool feature | Evidence | Added value | Burden | Decision | |---|---|---|---|---|---| ### Risk Matrix | Risk | Learner impact | Probability signal | Control | Owner | Residual risk | |---|---|---|---|---|---| ### Human Agency Matrix | Decision/task | Learner | Teacher | AI/tool | Human review required? | |---|---|---|---|---| ## 7. REQUIRED OUTPUT 1. Learning problem and objective. 2. Tool/AI fit analysis. 3. Evidence and limitations. 4. Privacy/accessibility/equity review. 5. Academic-integrity implications. 6. Human oversight. 7. Pilot/implementation plan. 8. Continue/modify/stop criteria. End with **EdTech & AI Integrity Check** confirming that technology serves learning rather than replacing the learner's or teacher's essential thinking. This supports educational technology use and does not replace local privacy, child-protection, accessibility or academic-integrity requirements.
分类与标签
学习教育、作业设计、教育人工智能、学术诚信、学习评量