✨Guide: Crafting Your Syllabus' AI Policy

Welcome to AutomatED: the newsletter on how to teach better with tech.

Each week, I share what I have learned — and am learning — about AI and tech in the university classroom. What works, what doesn't, and why.

In this fortnight’s Premium edition, I present a guide to help professors develop and express their overall stance on AI use in their courses. This guide covers both student use of AI and professorial use.

🖼️ The Big Picture

In this Guide to developing AI policies for university-level syllabi, I recommend that you, the professor, take a holistic approach that addresses both student AI use and your own professorial AI use. This not only creates an equitable culture of honesty and trust but also the opportunity for a dialogue between you and your students about the evolving role of AI in (their) education.

When you develop and refine your student AI use policy, I…

• recommend aligning your views on AI integration with institutional policies, departmental and field standards, and course objectives

• suggest that you need to consider how students would ideally use AI to achieve your courses’ objectives, before turning to incentivizing students to use AI in those ways (or, perhaps, incentivizing them to not use AI at all)

• discuss providing guidelines demarcating responsible AI use from AI misuse, as well as the importance of assessing students' familiarity with AI, ensuring equitable access, and providing AI training opportunities if needed

Regarding your development of your professorial AI use policy, I…

• advocate you start by clearly defining the scope of your AI use cases in connection with your course, ranging from brainstorming to using personally identifiable student data

• summarize some of the insights of our ✨Guide on ethically using AI with student data in discussing a range of data use strategies you might want to deploy, such as restricting AI use to non-student data or obtaining explicit consent when sensitive student data is involved

• stress the importance of transparency and flexibility depending on the context.

In both cases, I provide two generic examples of AI use policies, one AI-inclusive and one AI-exclusive, that incapsulate the lessons of the Guide and that can act as starter policies as you develop your own.

Developing Student AI Use Policy

The following steps offer a roadmap for you to create an AI policy for your syllabus that effectively manages AI's role in your students’ engagement with your classroom, ensuring that it aids in their learning rather than hinders it — and does not make your job more difficult.

This policy should align with institutional and departmental policies, respect the utility of AI use in your educational context, take into account accessibility issues with the relevant AI tools (including, if necessary, training opportunities), guide students on when and how AI tools can be responsibly used in the course, and set clear boundaries that demarcate how you will handle what you believe to be AI misuse.

Let’s cover each of these in turn.

👀 1 - Review Institutional Policies

The first step in developing an AI policy for a syllabus is to thoroughly review your institution’s policies, as well as those of your department. Universities and colleges often have detailed policies regarding technology use, academic integrity, and student conduct, which are essential to understand when integrating AI tools into your course.

I see these constraints — along with more general legal constraints — as guardrails that define the outer limits of your AI policies, regardless of your own views about what should happen within them.

Additionally, departmental guidelines can provide insights into discipline-specific considerations and best practices. You should consider whether your AI policy should be not only compliant with institutional rules but also resonate with the academic culture of your field and standards of your department, if relevant.

You can add language from both your institution and your department to your policy if you think it is necessary or if it would be helpful to your students.

✅ 2 - Consider Course Objectives
Relative to AI Use

The second step involves a careful examination of your course objectives in relation to the potential benefits and drawbacks of AI use. This step is key to determining how AI tools can align with or detract from the educational goals you have set. Begin by listing or reviewing the key learning outcomes or objectives of your course. Then, critically assess how AI tools can enhance or undermine these objectives.

We have covered aspects of this topic at greater length elsewhere, like in my ✨Guide on how to train students to use AI and my ✨Guide on how professors can discourage and prevent AI misuse, but broadly the two ends of the pedagogical spectrum are AI-inclusivity and AI-exclusivity…

This step requires relating the pedagogical opportunities and risks AI presents to the educational values you aim to uphold.

The goal is to determine the role that AI would ideally play in your course, such that you can define the terms of students’ engagement with AI to incentivize and encourage them to use it in that way. AI in your classroom should meaningfully contribute to the learning journey of your students, without compromising its integrity.

The result of this step won’t necessarily be any verbiage in your syllabus, but it will determine the contours of everything that you do write into your policy. If it is essential that students use a given AI tool, your policy needs to not preclude its use, account for your students’ varying abilities and access to the tool (see the next step), and so on.

🏋️ 3 - Assess Student Abilities and
Access, and Consider Training Options

Once you have figured out the role that AI would ideally play in your course — relative to your course’s specific objectives — you need to consider what you can expect of students.

Start by evaluating and estimating the current level of familiarity and proficiency your students have with AI technologies relevant to your course. This understanding is crucial as it influences how you integrate AI into your curriculum. For instance, in a majors-level computer science course, students will likely already have significant experience with coding assistants — perhaps they use GitHub Copilot regularly — whereas students in a first-year humanities course might not have used LLMs like Claude for drafting or ideation.

Next, assess the accessibility of AI tools for your students. This includes considerations of cost, technical requirements, availability, and accommodations.

Finally, you should consider how you can address any gaps in abilities through training and support. This could involve providing introductory sessions on AI tools, offering resources for self-learning, or incorporating AI tool tutorials into your course structure. Consider also the need for ongoing support, such as help desks or discussion forums, to assist students as they navigate and utilize AI tools throughout the course.

Ensuring equitable access is key; all students should have the opportunity to engage with the AI components of your course without burdensome barriers, whether financial, technological, or otherwise.

This step ensures that you are aware of what it will take to get your students equipped to engage with AI in a meaningful way, aligned with the course's learning objectives.

The results of these assessments, plans, and evaluations should be manifested in the language found in your syllabus. For instance, if you expect your students to use an AI tool that comes with a financial cost, you should consider whether you want to assign it as a requirement, like textbooks or other software. Whether this is reasonable depends on their means, expectations at your institution, the cost of the tool, and many other considerations. If, on the other hand, you determine that the cost of a very useful tool would be too great for some students, you need to address the fact that others do have the financial means, thereby granting them a potentially unfair advantage. This may require outright bans or other assignment structures that level the playing field.

🥕 4 - Reflect on Incentivization
and Accountability Frameworks

Developing an accountability framework for AI use in the classroom involves a cultural element first and foremost. Practical constraints are important but setting norms of ethical accountability are foundational.

For instance, you should consider the level of transparency you expect from students regarding their use of AI. This could involve requiring students to disclose their use of AI in assignments, submit chat logs (as recommended by AI for Education), or include reflective components in their submissions where they explain how and why they used AI tools in the ways they did. (In case you missed it, I discussed how to cite AI at some length earlier this year, including what the MLA and APA recommended.) These measures can foster a culture of openness and honesty in the classroom.

Next, it is important to think about practical methods for regulating AI use. While it is natural to think that the best strategy is to use software that can detect AI-generated content — and signal to students that it will be used in the syllabus and otherwise — there are reasons to worry about this strategy (consider the reservations I have expressed about these methods).

An alternative is to incorporate AI-resistant elements into your assignments. Design tasks that require personal reflection, in-depth analysis, or application of course-specific knowledge that AI tools cannot easily replicate. This approach not only challenges students to engage more deeply with the course material but also upholds the integrity of the learning process. Another strategy is using AutomatED's “pairing” method, which involves pairing AI-susceptible assignments with AI-immune ones to incentivize students to complete the former honestly and earnestly. For instance, you could assign a take-home essay that students then must defend in an oral exam. This method not only helps in maintaining academic integrity but also in assessing students' understanding and application of the course material independently of AI assistance.

With respect to your AI policy on your syllabus, the advantage of these alternatives is that, well, they do not require much of a policy! The structure of your course itself acts as the safeguard against AI misuse, rather than an outright ban or threat of detection.

⚖️ 5 - Reflect on Remediation
and Disciplinary Plan

Establishing clear guidelines and consequences for AI use is crucial. Develop and communicate guidelines that clearly outline what constitutes acceptable and unacceptable use of AI in your course. Pair these guidelines with specific consequences for misuse to ensure that students understand the boundaries and expectations set forth in your course.

(Note that a background assumption here is that you have provided support for students who might overly rely on AI due to gaps in understanding or skills. By providing these resources, you can help ensure that all students, regardless of their starting points, have the opportunity to succeed in the course without undue reliance on AI — and are not incentivized to misuse AI and violate your policy because you have not sufficiently prepared them for success. This should be addressed from step 3 above.)

Developing a disciplinary plan for AI policy violations requires a comprehensive and thoughtful approach. A graduated response to violations is wise. Begin with less severe consequences, such as warnings or the requirement to redo assignments for initial or minor infractions. Escalate to more significant consequences, like grade penalties, for serious or repeated violations. This approach ensures fairness and proportionality in response to various levels of infractions. It is also important to consider individual student circumstances when determining consequences. Factors such as intent, level of understanding, and previous record should influence the decision-making process, ensuring that each case is treated with due consideration.

In general, incorporating elements of restorative justice can be beneficial. Encourage students to reflect on their actions and understand the impact of their behavior. This approach promotes long-term understanding and respect for academic integrity and can be more effective than purely punitive measures.

Lastly, maintain a well-documented process for handling violations. Documentation is not only crucial for transparency and consistency but also for any necessary reviews or appeals. This ensures a clear and accountable process that upholds the integrity of your educational environment.

Transparency and consistency in the application of the disciplinary plan are key. Make sure students are well-informed about the guidelines and potential consequences at the start of the course by explicitly stating them in your syllabus AI policy.

🔬 6 - Implement, Get Feedback, Revise

Once the policy is integrated into your syllabus and classroom practices, it is vital to observe its impact and effectiveness. Solicit feedback from students to gauge how well the policy is understood, its effectiveness in addressing the concerns it was designed for, and its impact on student learning and engagement. This feedback can be gathered through surveys, open discussions, or informal feedback during office hours.

Creating a space for students to discuss their understanding of AI use and its implications can facilitate a deeper comprehension of the policy and its rationale, even if you end up unmoved with respect to making any changes to it in response to student feedback.

This dialogic stance is especially important given the range of perspectives students will have on AI use and misuse. Making students feel heard is key, not only because it requires you listening to them but it also requires them to articulate how they understand what you express with your policy.

Equally important is being open to revising the policy based on this feedback. It is unlikely that you will land on the perfect policy this semester (or the next, or the next, …). Adaptations might be needed to address unforeseen challenges, evolving AI capabilities, or changing student needs. This iterative process of implementation, feedback, and revision ensures that the AI policy remains relevant, effective, and aligned.

Examples

A Generic AI-Inclusive Example

A Generic AI-Exclusive Example

Developing Professorial AI Use Policy

Here’s my pitch for including a professorial AI use policy in your syllabus:

If you decide to include a professorial AI use policy in your syllabi, it should outline how AI tools will be utilized in your teaching and how you will handle student data, if necessary. The following steps provide a method to create a policy that makes room for you to leverage AI for educational enhancement but also mindful of the above ethical and cultural issues.

Subscribe to Premium to read the rest.