A person in a wheelchair works on a laptop at a desk in a bright office.

Before you ban, require, or ignore AI: A reflection framework for faculty

Best Practices, Teaching, Technology Trends, Tips

Many faculty are rethinking their learning activities, course policies, and assessments in light of generative AI. But the most useful starting point isn’t “How do I add AI to my course?” The better question is “What role, if any, should AI play in helping students meet the goals of my course?”

Faculty may feel pressure to respond to AI quickly, visibly, or innovatively. But thoughtful teaching sometimes means choosing not to use a new technology. A course doesn’t necessarily become more current or meaningful just because AI appears in an assignment. Instead, AI should be considered in relation to course goals, student learning, equity, academic integrity, and the kinds of thinking students need to practice in your course and/or discipline.

Faculty don’t have to integrate AI just because it’s available, popular, or administratively encouraged. But they do need to make deliberate choices about how AI intersects with their courses because students may already be using it.

However, even when faculty decide not to formally integrate AI, they still need to think about how students may already be using it. Faculty don’t have to integrate AI just because it’s available, popular, or administratively encouraged. But they do need to make deliberate choices about how AI intersects with their courses because students may already be using it (whether or not the syllabus mentions it).

The following reflective questions and framework are designed to help you decide how best to address AI in your courses.

What do I most want students to learn?

If AI shortcuts the process that students need to practice, it could interfere with learning. On the other hand, if it helps students access, extend, test, or reflect on their thinking, it could support learning.

Person pointing toward digital education icons around a light bulb and graduation cap above an open book.Some questions to consider:

  • What are the central habits of mind, skills, practices, or forms of knowledge students should develop in this course?
  • Which parts of the course require students to struggle productively?
  • Where do students need practice generating ideas, making judgments, revising, calculating, interpreting, creating, or explaining?
  • What kinds of student thinking do I need to see in order to assess their learning?

Where might AI interfere with the purpose of the task?

The same AI use can be harmful in one assignment and productive in another. The key is not whether AI appears, but whether it changes the intellectual work students are supposed to do.

Some points to consider:

  • Illustration of two AI-labeled computer chips in bubbles on opposite sides of a tilted balance scale, one with a halo and one with horns and a tail, against a gray world map background.If the purpose of a writing assignment is to help students practice developing an argument, having AI generate the argument may undermine the task.
  • If the purpose is revision, AI feedback may be useful, but students still need to evaluate and make decisions.
  • If the purpose is reading comprehension, AI summaries may prevent students from practicing sustained attention and interpretation.
  • If the purpose is coding fluency, AI-generated code may obscure whether students understand the logic.
  • If the purpose is critique, comparison, or verification, AI outputs may become useful objects of analysis.

Where might AI support learning?

AI shouldn’t be treated as inherently useful. It’s only useful when its use is aligned with a learning goal and when students are taught how to evaluate its output.

Illustration of a robot wearing a graduation cap and holding a green book in front of a yellow educational interface with icons and text lines.Possible beneficial uses may include:

  • Helping students brainstorm questions, not final answers.
  • Generating examples for students to critique.
  • Offering practice explanations that students must verify.
  • Supporting revision when students compare AI feedback with peer or instructor feedback.
  • Helping students identify assumptions, gaps, or counterarguments.
  • Simulating audiences, clients, patients, or stakeholders.
  • Supporting metacognition by asking students to explain how they used, rejected, or revised AI output.

How are students probably already using AI in this course?

Instead of treating all unsanctioned AI use simply as misconduct, faculty could ask what that use reveals about the course design, student motivation, workload, transparency, and assessment.

Person wearing glasses with glowing circuit-like patterns projected across their face and upper body, looking upward with one hand near their chin against a dark background.Some questions to consider:

  • Where are students most likely to use AI without permission?
  • Are they using it to avoid work, or to manage confusion, time pressure, language barriers, lack of confidence, or unclear expectations?
  • Which assignments are most vulnerable to outsourcing?
  • Where might AI use indicate that students don’t understand why the task matters?
  • What would students need to know in order to use AI more responsibly in this context?

What do I want students to understand about AI itself?

In some courses, the most meaningful AI integration may not involve students using AI to complete assignments. Instead, it might involve asking students to examine AI outputs critically.

Graphic with the text “AI makes decisions fast but not always right” above a person facing two paths, with smaller text reading “Bias risk,” “Wrong assumptions,” and “Flawed conclusions.”Examples of possible learning goals:

  • Students should understand that AI outputs can be fluent but inaccurate or specious.
  • Students should recognize that AI can reproduce bias, flatten complexity, invent sources, or present uncertainty as fact.
  • Students should understand that AI tools do not understand course material in the way human learners do.
  • Students should be able to verify claims, trace sources, compare outputs, and explain where AI reasoning fails.
  • Students should understand privacy, data, authorship, intellectual property, and disciplinary norms.

What use is allowed, required, discouraged, or prohibited?

In addition to explaining the parameters of AI use for your course, you should also be very specific about what those parameters involve. Avoid vague language like “use AI responsibly” unless you explain what responsible use means in that specific course.

Person wearing magenta sneakers standing at a three-way choice marked “Yes,” “Maybe,” and “No” on a textured ground surface.Example policy framework:

  • Allowed: Students may use AI for brainstorming, study questions, grammar support, or revision suggestions.
  • Required: Students must use AI as part of a specific assignment and analyze its limitations.
  • Discouraged: Students should avoid AI in early drafting, problem-solving, or reading responses because those activities are intended to build foundational skills.
  • Prohibited: Students may not submit AI-generated work as their own or use AI on assessments designed to measure unaided performance.

What does transparency look like?

Transparency shouldn’t just be framed as surveillance (which can erode student trust). It can also support metacognition by asking students to explain their process and choices.

Person in a light blue sweater looking upward with one hand near their chin against a light gray background.Some questions to consider:

  • Should students disclose whether they used AI? (Also, should you disclose when you use AI?)
  • Should they name the tool?
  • Should they include prompts?
  • Should they describe what they accepted, rejected, changed, or verified?
  • Should they submit an AI-use reflection?
  • Should disclosure be graded, required but ungraded, or simply part of academic integrity expectations?

How will I assess learning since AI is available?

AI is widely available to students, and no assessment design is fully “AI-proof.” A more productive approach is to ask what evidence of learning an assignment produces. Does it reveal students’ reasoning, their process, their ability to apply course concepts, their judgment, or only their ability to submit a polished final product? What kinds of evidence would show that students are developing the knowledge, skills, and judgment the assignment is meant to cultivate?

Graphic of a head outline filled with gear shapes and a paper label reading “Learning Outcomes” in red text.Some options:

  • Ask for process artifacts: notes, drafts, annotations, revision plans, lab notebooks, decision logs.
  • Use in-class writing, oral explanation, demonstrations, or conferences when appropriate and feasible.
  • Design assignments that require course-specific, local, personal, experiential, or contextual analysis.
  • Ask students to critique AI-generated responses using course concepts.
  • Assess judgment, justification, reflection, and transfer rather than only polished final products.
  • Use scaffolded assignments where students show development over time.

What equity issues should I consider?

Equity doesn’t automatically mean allowing AI, and rigor does not automatically mean banning it. Ultimately, faculty need policies that are clear, teachable, and consistently applied.

Icon of two hands holding a group of people above a heart symbol.A few equity considerations:

  • Students may have unequal access to paid AI tools.
    If AI use is required or encouraged, will I provide access to a common tool, offer a no-AI alternative, or design the task so paid-tool advantages do not determine success? How will I avoid rewarding students simply because they have access to more powerful tools?
  • Students may differ in prior experience with AI.
    What minimal guidance would students need in order to understand acceptable and unacceptable uses in this course? How can I make expectations explicit enough that students are not advantaged or penalized based on prior AI experience?
  • Multilingual students may use AI for language support in ways that complicate simple authorship policies.
    What kinds of language support are appropriate in this course, and where would that support begin to interfere with the learning goals?
  • Students with disabilities may experience AI as an accessibility support.
    If I restrict AI use, have I clarified how students with accommodations should navigate that policy? Are there non-AI alternatives that provide equivalent access without requiring students to disclose private information unnecessarily?
  • Surveillance-oriented responses may disproportionately affect students already subject to mistrust.
    Could my response to AI use increase mistrust between students and me? Who is most likely to be questioned, scrutinized, or accused under this policy? What evidence would I need before treating a case as misconduct? Can I design assignments and disclosure practices that emphasize learning, transparency, and process rather than surveillance?
  • A total ban may be difficult to enforce consistently.
    If I say AI is not allowed, how will I communicate what that means in practical terms? Are there parts of the course where a supervised, in-class, oral, process-based, or scaffolded assessment would better align with my concerns? What will I do if I suspect AI use but cannot prove it?

What is my disciplinary context?

Illustration of colorful textbooks labeled with school subjects including geometry, history, physics, biology, literature, chemistry, English, geography, mathematics, and philosophy.AI use in a composition course, clinical course, programming course, design studio, lab science course, or history seminar may require very different policies and pedagogical choices.

Some questions you may want to consider as you develop your AI policies and instructional decisions:

  • How is AI being used, debated, or restricted in my field?
  • What forms of expertise do students need to develop before AI use is appropriate?
  • What professional standards, ethical norms, or accreditation requirements matter?
  • What would responsible AI use look like in this discipline?
  • What AI-related mistakes would be especially consequential?

Final thoughts

Faculty don’t have to make AI central to their courses. However, they do need to decide how AI relates to the course’s learning goals, assignments, assessments, and students’ actual practices. That decision might mean stipulating that AI use isn’t appropriate for certain tasks, particularly when the purpose of the task is to practice foundational thinking, reading, writing, problem-solving, or disciplinary judgment.

[A] course policy alone isn’t going to make students responsible AI users.

Yet, a course policy alone isn’t going to make students responsible AI users. Even when faculty don’t invite AI into their assignments (or prohibit its use), students still need guidance on what the tools can and cannot do, why certain uses undermine their learning (and could negatively impact their ability to do more complex learning in future courses), and how to evaluate AI-generated outputs critically. A meaningful faculty response to AI isn’t just a question of permission or prohibition. It’s a question of pedagogy.

Illustration of several people moving around and interacting with a large purple brain, books, papers, and abstract ribbon-like shapes on a dark background.

In some courses, responsible pedagogy may mean asking students to analyze flawed or specious AI outputs, compare AI-generated claims against course readings, identify hallucinated or weak evidence, or reflect on when using an AI tool becomes a substitute for real learning.

In other courses, it may mean explaining why students are being asked to work without AI for a particular assignment or what skills or intellectual capacities that task is supposed to build. The goal isn’t to pretend that AI can be kept out of the learning environment entirely (it can’t). Rather, the goal is helping students understand the difference between using AI as a tool and outsourcing their own thinking or accepting machine-generated output uncritically.

How can we help students develop the judgment they need in contexts where AI is available, unevenly reliable, and easy to misuse?

A thoughtful AI policy is part of a larger teaching decision rather than just an exercise in boundary setting. How can we help students develop the judgment they need in contexts where AI is available, unevenly reliable, and easy to misuse? Faculty can decide not to allow AI for particular coursework or the course as a whole, even if such policies are difficult if not impossible to enforce; at the same time, they can help students become more critical and responsible AI users.

Resources

NIU:

Outside Resources:

Scholarship:

Print Friendly, PDF & Email