Explore AI Fundamentals

Clear and Transparent AI Guidance

Having a clearly articulated AI policy in your class is good for students, good for faculty, and helps everyone better adhere to academic integrity principles. Proactively set a clear AI policy about what is allowed and not allowed in your class. Include this guidance in your syllabus and discuss it with your students regularly, particularly before big assignments are due. Highlighting the policy frequently helps frame AI use as an ongoing conversation about academic integrity rather than a one-time rule.

When designing your AI guidance for your students, please try to be as clear and specific as possible when discussing appropriate and inappropriate AI use. Explore examples of this language on our “Developing AI Syllabus Statements” page.

Define what is allowed and what is not allowed.


Be explicit about what counts as AI misuse. For example, you may want to tell your students: “Using Gemini to outline your essay is not permitted unless specifically stated. However, using Grammarly for copy-editing and proofreading is generally allowed.” 

1. Leverage the AI Assessment Scale (AIAS)


Use a framework like the five-level
AI Assessment Scale (AIAS) to clearly indicate the extent to which students can use AI on assignments. For example, with the AIAS framework (Perkins, et al, 2024), you can designate assignments with one of the following levels:

 

  • No AI: Assignments are to be completed entirely without AI assistance.
  • AI Planning: Students can use AI for brainstorming, outlining, and generating new ideas. Students are tasked with developing and refining the ideas generated by AI.
  • AI Collaboration: Students can use AI for revisions and feedback to improve the clarity and quality of their work. Students are tasked with critically evaluating the AI generated output.
  • Full AI: Students can use AI extensively to complete any aspect of the assignment.
  • AI Exploration: Students use AI creatively to develop innovative approaches and novel questions.

2. Create a Skill-Based AI Permission Table


Use a table to define which tasks, tools, or skills are relevant for your course. Then, indicate whether student use of AI assistance is prohibited or permitted for each skill, using the following three options:

  • Permitted without disclosure: Students can use an AI tool to accomplish this skill
  • Permitted with disclosure: Students may use an AI tool to accomplish this skill as long as they tell you about it
  • Prohibited: Students are not allowed to use an AI tool to accomplish this skill
Template skill-based AI permission table
Skill AI Assistance is:
Freely permitted with no need to disclose Permitted with disclosure Prohibited
Grammarly or similar/style checkers
AI research assistants (e.g., Perplexity)
Brainstorming or outlining with AI
Summarizing texts with AI
Generating full drafts with AI
Using AI to critique their own writing
Translation of non-English resources
Optional character recognition (OCR)
Transcription (handwritten/audio)
Text-as-data analysis (e.g., topic modeling)

Bibliography

Looking for more resources on AI use in higher education?

We invite Columbia University faculty and graduate students to connect with the Center for Teaching and Learning (CTL) team to discuss how AI can be used purposefully and ethically in higher education. Schedule an in-person consultation with our team, visit our open Office Hours, or log in for our virtual chats to explore which AI tools align with your teaching goals and how to implement them effectively.