Framework for Teaching Judgment in the Use of AI

The genesis of the REACT Framework

When generative AI first appeared in our classrooms, many schools were filled with concern about having to deal with uncertainty. Everywhere, teachers faced the same question: what to do with this powerful and unpredictable new technology? 

The instinct to control or ban the use of AI is understandable, but it is not a real solution. Banning AI in schools would only delay the inevitable and would not prepare our students for the world they are about to face. 

Between blocking AI and its full adoption, Morris had already introduced an essential intermediate step at the end of 2024: reflection questions that helped students examine how and why they were using AI. This approach laid the foundation for what became the REACT Framework, shifting the focus from compliance to reflection and, ultimately to professional practice. 

React! 

We recall a conversation between us where, at one point, almost without thinking, we both said the same thing out loud: “We can’t just ban AI. We have to at least react!” 

This simple statement became the starting point for reflection. React! But how? 

We started with 2 very concrete questions that every teacher is asking themselves today: 

• How can we fairly evaluate work when, in many cases, students will use AI at some point in their process? 
• How can we ensure that students are really learning and developing their judgment and ethical sense, despite (and sometimes thanks to) this use? 

These questions force us to reflect beyond the “allow or prohibit” dilemma and turn it into a pedagogical design issue: creating a system that makes reasoning visible, not just the deliverable. This system is the REACT Framework. What began as a conversation between 2 teachers has since become a flexible framework that educators can adapt, experiment with, and make their own. 

In designing the REACT Framework, professional practice was our “ideal scenario.” If students could learn during their studies to use AI judiciously, responsibly, and professionally, we would be giving them a gift that would serve them well beyond the classroom. 

The evolution of AI use: from compliance to reflection to professional practice 

As we mentioned, when AI first appeared in our lives, many people’s reaction was to adopt a compliance approach: policies focused on prohibition and disclosure were put in place. The advantage of this approach is that it establishes safeguards. However, it promotes minimal engagement by encouraging a “checkbox mentality.” 

Others have adopted a reflective approach by relying on AI reflection journals. The advantage of this approach is that it encourages metacognition and transparency. However, it can lead to inconsistencies. Furthermore, when used without a framework, a reflection journal is unfortunately often considered a supplement rather than an essential part of the decision-making process. 

Our approach, the REACT approach, is focused on professional practice. It integrates compliance and reflection into a competency-based system. Students are assessed on process, accountability, and professional maturity, not just on results. 

REACT goes beyond structured thinking; it supports professional practice by preparing the students to: 

• apply AI with judgment, without dependence 
• take ownership of their decisions and demonstrate a sense of ethical responsibility 
• clearly express the value of human beings in complex workflows 
• move from being users of AI to professionals who master AI 

This is not a replacement for disclosure policies regarding the use of AI, but rather their natural evolution. 

The REACT Framework 

We created the REACT Framework based on the key skills required in the job market for 2025-2030. To identify these skills, we drew on numerous sources, such as the World Economic Forum’s The Future of Jobs Report 2025. 

During our research, we grouped the skills into what we called REACT, a thematic structure designed to address the real challenges of working between humans and AI.  

The 7 assessment questions of the REACT Framework 

The REACT Framework is based on 7 questions. 

These questions take on different meanings depending on when they are asked. At the beginning of a course, they serve best as formative tools that encourage self-reflection and autonomy. Later on, they become summative tools that act as markers of responsibility in professional practice. 

In a formative assessment context (during learning) 

In the 1^st part of the semester, to teach our students how to use AI, we have them keep a reflection journal on professional practice in which they must answer 7 questions.  

This learning activity helps students think critically before the final evaluation. Rather than testing results, the goal is to create a feedback loop between students and us. Students reflect, we respond, and in this way, everyone adapts and learns.  

The reflection journal is intertwined with individual case studies, essays, papers, articles, and reports. It supports assignments that require students to express their reasoning, decision making, and ethical awareness in their work. (The journal is not intended for multiple-choice tests, quizzes, or graded group assignments, as these do not allow for meaningful individual reflection or professional insight.) 

At the beginning, students are invited to reflect on simple yet powerful questions: “Should I use AI here? If so, why? If not, why? What are the issues at stake?” 

This step is devoted to exploration.  

Students have 2 options: 

• if they choose to use AI, they must say what they asked the AI, present the answers they obtained, and explain their decision to keep or reject certain information. 
• if they choose not to use AI, they must explain the alternative strategies they used. 

Both options are valid. What matters is the reasoning. This formative work paves the way for feedback: teachers can review the reflections and then discuss them with the students. This is where the feedback loop begins: a space for dialogue on action, ethics, and early decision making. 

Here are the 7 questions from the REACT Framework applied in a formative assessment context. Each question prompts a specific type of reflection and provides an opportunity for teacher feedback.  

In a summative assessment context (demonstrating professional practice) 

Towards the end of the semester, when students are working on major projects or final assignments, the same 7 questions reappear, but from a different angle. Instead of “Should I use AI?“, the question becomes “How did I use it (or not) and what does that reveal about my professional practice?“ 

Students once again have the power to act. They can: 

• fully integrate AI, demonstrating responsible use, verifying results, and adding unique human value 
• work without using AI, while reflecting on why this choice aligns with their professional identity and how they might interact differently with AI in a professional setting 

The REACT reference framework template [docx] includes a detailed evaluation grid.

What students say about the REACT approach (based on their reflections in class)

Here are some anonymous and representative quotes, provided with the students’ permission. Since few students chose not to use AI, we have a more limited number of comments to share.

Quotes from a student who would not use AI

We didn’t use AI because market analysis is a task that requires human judgment … we don’t want a tool to provide us with “relevant” information that could mislead our company.

Quotes from students who would use AI

If I used AI, I would seek to save time and organize my ideas efficiently.

If I used AI … I would apply it to analyze trends in the healthcare supply chain more effectively.

We decided to use AI for proofreading and verification to eliminate human error.

I would use AI because it allows you to analyze large amounts of data to predict future costs during financial planning.

If I planned to use AI, I would provide detailed context, specify constraints, and require clear citations and boundaries for acceptable content.

Potential for teachers

For teachers, adopting REACT is not so much about following a model as it is about participating in a conversation:

• How can we teach responsibility in a positive context?
• How can we design learning experiences that respect privacy, consent, and creativity? 
• How can we prepare students to make ethical choices in real-world contexts, not just in theory? 

Each teacher will find a different way to answer these questions, depending on their discipline and courses. REACT was designed precisely for this diversity. It is a framework for exploration, adaptable, and capable of integrating into any teaching style in any discipline. 

We have observed that when our students engage with these 7 questions, something important happens: they begin to see themselves as professionals in training, rather than just learners completing a specific task. Students practise being responsible, transparent, and reflective. They adopt the character traits that higher education strives to develop! 

If the goal of education is not limited to imparting knowledge, but also to shaping judgment, then reference frameworks such as REACT remind us that our job is not to provide all the answers, but to help students learn to ask better questions and to do so responsibly, in partnership with the technologies that will define their world. 

A colleague from Champlain College Saint-Lambert, Stéphane Paquet, is enriching our educational ecosystem by developing catalogues of AI prompts designed for the CEGEP leveI. Thanks to this work, it is becoming easier for us to build learning activities that remain fully consistent with the spirit and structure of REACT. 

When AI takes over: validation meeting  

If an assignment does not seem to reflect a student’s reasoning or does not correspond to what they normally demonstrate in class, and if the tone of the answers to the 7 REACT questions seems generated rather than thoughtful, a short 10-minute validation meeting is a legitimate practice. It should be announced in the course outline and presented as a learning support, not as a punishment. Transparency sets expectations, reduces anxiety, and maintains a positive, learning-centered environment. We invite the student in by explaining the goal: to confirm their understanding of the topic and acknowledge the personal aspect of their learning process.  

The conversation should be simple, focusing on 2 quick questions:  

• How did you verify the facts, your sources, their originality, adherence to guidelines and confidentiality, etc.?  
• What is your own contribution (your words, your decisions, what you actually changed or added)?  

At the end of the meeting, provide concise feedback. If the understanding is clear, the grade is confirmed; otherwise, the student is asked to redo the work in order to deepen their knowledge, in the most relevant format (for example, redoing the REACT reflection in the “with AI” annotated or “without AI” path, strengthening verifications and references, or briefly presenting the approach orally).  

The intention remains the same: to refocus the activity on learning and deepening knowledge. 

Potential for work and real life 

The REACT Framework is already being implemented at Champlain College Saint-Lambert (regular, daytime, full-time programs) and at York University in continuing education courses to guide real decisions in coursework, assignments, and teacher evaluation processes. 

Beyond the classroom, REACT naturally applies to everyday business life. For example, human resources management teams can use REACT to: 

• structure quarterly and annual evaluations 
• calibrate expectations 
• document how evidence is collected and verified 

Anyone can use it as a simple mental model to decide when and how to integrate AI into their daily tasks. In short, we plan to offer multiple practical applications across both education and the workplace. 

Big news for 2026 

We will definitely be using REACT throughout 2025-2026. In fact, at the time of publication, Champlain College Saint-Lambert announced that in 2026 it will launch the 1^st  technical program (DSC) in Business Administration in Quebec that fully integrates AI into business. 

Let’s share our practices 

If the REACT Framework seems useful in your courses or program, use it! Adapt it to your needs if necessary, and let us know how it goes! 

We have written this text as an invitation, not an obligation. If you are a teacher, program coordinator, or administrator exploring AI in your courses, we would love to hear from you. Share your own methods and ideas for teaching with or without AI. If you plan to try the REACT Framework, or if you have already used it, tell us what you did and what you observed. Get in touch!