ChatGPT, Codex, and AI

Practical AI workflows for faculty

Devin De Frisco
Devin De Frisco
  • Updated

Good first use: Ask AI to transform materials you already trust—such as turning your own lecture notes into practice questions—then review the result for accuracy, accessibility, and alignment with your course.

AI is most useful when it reduces drafting and formatting work while you retain the academic judgment. The examples below are starting points, not substitutes for course expertise or required instructional processes.

Choose a workflow

  • Create a structured first draft

    Using the attached course objective and my notes, draft a 50-minute lesson plan.
    Include an opening question, a 15-minute mini-lesson, a small-group activity,
    three debrief questions, and a two-minute exit ticket.
    Identify where students may need prerequisite knowledge.
    Do not introduce facts that are not supported by the attached materials.
  • Generate low-stakes practice

    Use the attached reading to create eight practice questions for students.
    Mix recall, application, and analysis. Include an answer key with a one-sentence
    explanation and the relevant page number. Flag any question whose answer is ambiguous.
  • Find barriers for human review

    Review this handout for accessibility and clarity. Identify problems with headings,
    link text, table structure, reading order, color-dependent instructions, and jargon.
    Propose revisions, but keep technical terms that students must learn.
    Do not claim that the document is compliant; give me a checklist to verify.
  • Turn criteria into observable language

    Draft an analytic rubric for the attached assignment and learning outcomes.
    Use four performance levels and observable criteria. Avoid vague words such as
    “good” or “excellent” unless the descriptor explains what they mean.
    After the table, identify possible bias or accessibility concerns for my review.

Faculty review checklist

  • The content matches the course learning outcomes and current disciplinary knowledge.
  • Examples and questions have correct, unambiguous answers.
  • The difficulty, workload, and reading level are appropriate.
  • Materials do not introduce stereotypes, fabricated sources, or misleading quotations.
  • Documents and activities support accessibility and approved accommodations.
  • Your use follows department, school, copyright, privacy, and academic-integrity expectations.

Be transparent with students

When AI materially shapes course content, assessment, or feedback, use the disclosure practice required by SBS or your academic unit. Tell students what is permitted for an assignment, what must be their own work, and how they should acknowledge AI assistance.

Student information: Do not submit identifiable student records, grades, accommodation information, advising details, or other protected information unless the specific tool, account, and use are approved under SBS policy.

Related guides

Official guidance

See OpenAI’s Prompting and Work with files guides. SBS academic and information-handling policies take precedence.

Related to