Generative AI (GenAI) has been stirring up a lot of discussion in higher education circles lately with the advent and availability of emerging AI tools like CoPilot, ChatGPT, and others. Understandably, many university faculty have expressed concerns about how AI tools might impact academic integrity or even lead to new forms of plagiarism if left unchecked. These concerns are legitimate and getting due consideration across campuses nationwide.

At the same time, what if we allowed ourselves to investigate these tools from a different vantage point? How might leveraging GenAI make faculty’s teaching lives easier? In particular, this may be most useful when assisting faculty in accomplishing some of the most time-consuming but high-impact teaching tasks, like providing meaningful and timely feedback to students, including in online, hybrid, and multi-modal learning contexts.

Let’s explore how AI-powered tools, especially those based on Large Language Models (LLMs), could potentially help faculty save time, deliver more personalized feedback, and keep everything above board regarding student privacy.

Student Privacy and AI Concerns

Any AI platform used for student work should be in a ‘walled garden’, meaning that students’ Intellectual Property is protected and their data remains secure. Some institutions prohibit instructor use of GenAI for grading and giving feedback. We recommend that faculty always check their institution’s policies regarding GenAI before using any tool in their teaching practice. Boston College has invested in Microsoft Copilot and worked to ensure that none of the content entered into the platform will be used for training purposes. 

I make the case that no summative assessments should be reviewed by AI. Rather, AI can assist with the challenge of providing continuous formative feedback for students. Faculty often face structural constraints, such as increasing class sizes, which might mean that providing their students with the type of specific and timely feedback they need to best support their learning is difficult or impossible. In these situations, faculty might need to weigh their options, determining whether providing little or cursory feedback versus leveraging a tool like GenAI to provide more specific, personalized feedback that best fits their instructional needs.  Here, faculty members need to decide if the ethics of not giving students feedback to support their learning outweigh any concerns they may have about submitting student work to AI. If the decision is to move forward with this, then, of course, faculty should clearly communicate when and how they will use AI and allow students to opt-out. 

What is a Prompt, and How Does It Work?

At the heart of using AI tools like CoPilot, Gemini, ChatGPT, or other GenAI platforms is a ‘prompt.’ A prompt is a set of instructions or questions you give the AI to shape and direct its response. Think of it as a way of asking the AI to perform a specific task, much like you might ask a teaching assistant. These tasks include evaluating the mechanics of an essay, offering customized feedback based on a rubric, or generating suggestions for improvement. The prompt tells the AI what you want it to focus on, how detailed the response should be, the directive’s context, and the answer’s format.

The magic of AI-generated feedback lies in its prompt engineering—this is the practice of carefully crafting these prompts to get the most valuable output from the AI. The better the prompt, the more relevant and helpful the AI’s response will be. For example, asking the AI to ‘Evaluate the critical thinking displayed in this essay’ will get a much more nuanced answer than simply saying, “Give the student some feedback on this paper.’ The more precise you are with your prompts, the better the AI will understand what you need.

Useful Tool or New Cheating Tool? Finding the Balance with AI

It’s no secret that some instructors worry about AI tools like ChatGPT being used by students to cut corners on assignments. That’s a valid concern, especially with how fast these technologies are evolving. But here’s the thing—AI can also be a powerful ally in your teaching toolkit. It’s not about replacing your expertise; it’s about automating the repetitive tasks that eat up your time so you can focus on what you do best—guiding and inspiring your students.

While AI tools like ChatGPT can generate clear and detailed feedback, the trick is how we use them. The aim should be to position AI as a support system for both instructors and students, allowing you to focus more on strategic guidance and less on the minutiae of grading. 

How LLMs Can Enhance Feedback

If you’ve ever found yourself drowning in a pile of essays to grade, you know how overwhelming it can be to provide individualized feedback to each student. AI-powered tools can assist by taking on some of that workload while still letting you personalize the feedback. Studies have shown that LLMs like GPT-4 can be used to evaluate student submissions using predefined rubrics, offering consistent, structured feedback that lines up with your course objectives. 

Practical Prompts to Get You Started

Let’s make this even more actionable with some practical prompts you can use with AI tools like CoPilot or ChatGPT to smooth the grading and feedback process. You can use these prompts to guide the AI in providing feedback that’s not just automated but also meaningful.  In each instance, you’d want to provide context regarding the specific requirements of the assignment/assessment, and information about your expectations, including any evaluation criteria as well as any guidance or success criteria you provided to the learners: 

Basic Feedback Prompt:

Provide feedback on this student’s essay focusing on clarity, coherence, and argument strength. Highlight areas where the student excelled and areas that need improvement, specifically in terms of supporting evidence and logical flow.

Rubric-Based Feedback Prompt:

Using the following rubric, grade the student’s essay:

• Thesis Statement: Does it clearly state the argument? (0-5 points)
• Evidence and Support: How well does the student use evidence to back up their argument? (0-5 points)
• Organization and Structure: Is the essay well-organized with a logical flow? (0-5 points)
• Grammar and Style: Are there any major grammar or style issues? (0-5 points)

Provide detailed feedback on each category, and suggest specific ways the student can improve.

Supportive Feedback Prompt:

Analyze this essay and provide feedback as if you were the student’s peer, using supportive and encouraging language. Focus on one positive aspect of their work before pointing out one area where they can improve.

Encouraging Critical Thinking:

Evaluate the critical thinking displayed in this essay. Does the student question assumptions, make logical connections, and consider multiple perspectives? Provide feedback on how they can deepen their analysis in future assignments.

Best Practices for AI-Enhanced Feedback

1. Experiment and Optimize: Don’t be afraid to try different prompts to see which ones yield the best results. AI is most effective when guided by clear and specific instructions.
2. Combine AI Feedback with Your Insights: See AI  as your co-pilot. Use it to generate a first draft of feedback, and then personalize it with your own insights to make the comments even more relevant and specific to each student’s work.
3. Teach AI Literacy to Students: Encourage students to use AI-generated feedback as a learning tool rather than a shortcut. Helping them understand how to interpret and critically evaluate this feedback can turn AI into a powerful educational resource rather than just a quick fix.

The Importance of Transparency with Students

As you integrate GenAI tools into the feedback and grading process, it’s crucial to be transparent with students about how you will use them in your class. Clearly communicating that you’re using AI to support the evaluation process sets clear expectations and fosters trust in the learning environment. This transparency also helps demystify these technologies, encouraging students to view them as valuable tools rather than shortcuts or replacements for human effort. One effective way to do this is by including an AI usage statement in your syllabus that outlines how you’ll be using the technology, as well as any limitations or expectations you have regarding their potential use of GenAI.

Disclosing your use of GenAI also provides an opportunity to explain how these tools give you more time to focus on providing individualized instruction, coaching, and support. By automating the repetitive aspects of grading, AI allows you to engage with students on a deeper level, offering personalized guidance, context, answers to their questions, and addressing their unique learning needs. This approach not only enhances the overall learning experience but also strengthens the connections you build with your students.

Ultimately, the goal is to use AI to enhance, not replace, the instructor-student relationship. When students understand that AI helps you to concentrate more on their growth and success, they’re more likely to see this technology as a positive force in their learning experience. Being open about GenAI’s role in assisting with the feedback process reinforces the idea that you are there to guide and support them every step of the way, using the latest tools to maximize their learning outcomes.

Wrapping It Up: AI as Your Teaching Assistant

The future of AI in education isn’t about removing the human element but enhancing it. While it’s natural to have concerns about the impact of AI on academic integrity, there’s also a massive opportunity here to make your teaching tasks more manageable and impactful. AI tools can streamline the feedback process, allowing you to focus on personalized interactions with your students without compromising quality or fairness.

Let’s reframe the conversation—AI doesn’t have to be the enemy of academia. Instead, it can be your ally in delivering high-quality learning experiences that are both efficient and tailored to the needs of each student. With a thoughtful approach, AI could potentially help you turn the task of providing feedback into an opportunity for deeper student engagement and learning and more high-level guidance based on your knowledge of the subject matter, and focus less on the more tedious and repetitive tasks of creating logistical, grammatical, and structural feedback for each learner on each assignment. 

For more insights on effectively implementing AI and other innovative technologies in your teaching practices, keep an eye on the CDIL Resources within the Digital Teaching Toolkit.