The Power of a Pause: Embedding Ethical Guidance in Student-AI Interaction

Strategic Minds AI

Advithiya More, Pace University(NY) Milagros Espejo Bocanegra, Pace University(NY) Esuona Sulo, Pace University(NY) Brendon Khoza, Pace University(NY) Joshua Kolawole, Pace University(NY)

Training / education / pedagogy

This project explores a scenario-based AI guidance system designed for students in early academic stages. Through a simulated interface, students engage with realistic academic situations (e.g., essays, assignments, exams) and receive AI assistance either with or without structured guidance on rules and ethical use. The study evaluates how embedded guidance influences students’ interpretation of acceptable AI practices in higher education.

AI adoption in higher education has rapidly outpaced institutional guidance, creating a gap between widespread use and clear governance. While concerns around academic integrity are significant, an equally critical issue is the potential decline in meaningful learning when students rely on AI to complete academic tasks without cognitive engagement. This pattern of use can undermine critical thinking, impede the development of writing skills, and weaken long-term knowledge retention . At the same time, unclear policies lead to inconsistent practices and inequities among students. This project addresses both ethical and learning risks by testing whether structured, embedded AI guidance can promote responsible use while encouraging active participation in the learning process

Our team is motivated to explore how AI can be integrated into education in a way that supports not only ethical use, but also meaningful learning. We aim to move beyond performance-driven outcomes and examine how students make decisions when using AI in real academic contexts. Specifically, we are interested in how guidance can encourage active engagement, critical thinking, and skill development, rather than passive reliance on AI. Through this project, we will learn how students interpret AI boundaries, how guidance shapes behavior, and how to design AI systems that embed ethics, governance, and learning support—contributing to more human-centered, transparent, and accountable educational technologies.

1. What is the situation or context you are addressing? The project addresses the growing use of AI tools in higher education, particularly among early-stage students. Students are increasingly using AI for academic tasks such as writing essays and solving assignments, often without a clear understanding of what constitutes acceptable use. This rapid adoption has outpaced the development of clear institutional policies and guidance defining what constitutes appropriate or ethical use of AI. As a result, students are often left to interpret these boundaries on their own, leading to inconsistent practices. Some may use AI as a support tool to enhance understanding, while others may rely on it to complete tasks with minimal cognitive effort. This situation raises not only ethical concerns but also potential risks to the learning process itself, particularly in the development of critical thinking and writing skills. Consider a scenario where a student has 3 hours before an assignment submission deadline. The student is well versed in the topic, has completed the readings and has genuine ideas related to the topic. However, time has become her biggest constraint. She opens an AI interface and prompts it to write an introductory paragraph. Seconds later, the AI responds and she proceeds to paste it in without second thought, not because she decided to cheat but nothing in that exact moment prompted her to stop and to choose consciously. No rule or prompt asks her to reflect at that exact moment of decision. The rules and boundaries between AI assistance and AI substitution existed somewhere in the university policy documents but not where it mattered the most: at the precise and exact moment of decision. 2. What is your critical analysis of this situation? The situation reflects a critical misalignment between technology adoption and educational governance. While AI offers clear benefits in terms of accessibility and efficiency, it’s unstructured use risks undermining core learning processes. First, over-reliance on AI, using it as a substitute and not as a support tool, can reduce cognitive engagement, limit the development of critical thinking and writing skills, and negatively impact long-term knowledge retention. At the same time, the absence of clear, embedded guidance shifts responsibility entirely onto students, who may lack the experience to navigate these boundaries. This not only in ethical concerns but also in inequities, as students with better informal knowledge of AI practices gain an advantage over others. Therefore, the issue is not simply misuse, but the lack of structured systems that actively guide responsible and pedagogically sound AI interaction. 3. What perspectives were discussed and how were they debated or arbitrated within your team? The team considered multiple perspectives when discussing the role of AI in higher education. One perspective emphasized academic integrity, focusing on preventing misuse and maintaining fairness in assessment. Another view highlighted the importance of learning outcomes, arguing that even “allowed” AI use could be harmful if it replaces active thinking, weakens engagement, and reduces long-term retention. A third perspective focused on student autonomy, suggesting that overly restrictive systems might limit exploration, innovation and the opportunity to learn how to use AI responsibly. These perspectives were debated taking in consideration trade-offs between control and flexibility. As students ourselves, an important part of the discussion was the need for clarity. We want to be sure that we are not violating university rules, but at the same time, we also recognize that AI has already become part of academic life. We also acknowledged a practical reality: AI use is becoming almost unavoidable as students are competing in environments where others are already using it. Ignoring this reality or attempting to ban AI altogether does not seem realistic or productive. However, a purely permissive approach without any structure could reinforce misuse and create unfair advantages for students who already know how to navigate these tools more effectively. The team ultimately aligned on a balanced approach: instead of restricting AI use, the system should make decision-making more transparent. By embedding guidance at the point of interaction, the solution preserves student agency while encouraging informed and responsible choices. This arbitration reflects a shift from enforcement-based models to guidance-based design, not only addresses ethical concerns, but also aligns AI use with meaningful learning objectives. 4. What contribution are you proposing, and under what conditions could it be implemented? Our contribution is a scenario-based AI guidance system that embeds structured prompts within student–AI interactions. Instead of passively generating responses, the AI introduces decision checkpoints that encourage students to reflect on how they want to use AI (e.g., idea support vs. content generation). This transforms AI from a task executor into a learning-aware guide, making ethical boundaries visible at the exact moment of use and promoting more intentional engagement. Experimental Design (Control vs Experimental Conditions) To evaluate the effectiveness of this system, we propose a controlled experimental study with two conditions: 1. Control Group (No Guidance): - Students interact with AI freely - AI responds directly to prompts without intervention 2. Experimental Group (Embedded Guidance): - AI introduces structured prompts before responding - Example: “Do you want help refining your ideas or generating content? Your course policy allows the first but not the second.” What will be measured? (Key Variables): - We aim to capture both behavioral and cognitive outcomes: 1. Decision-Making Behavior - Type of AI use (generation vs support) - Frequency of AI reliance - Changes in choices after guidance prompts 2. Ethical Understanding: - Awareness of institutional policies - Ability to distinguish between ethical vs unethical usage 3. Learning Engagement: - Depth of student-written content - Evidence of original thinking vs AI dependency - Time spent actively engaging with the task 4. Output Quality: - Essay coherence, argument strength, and originality - Comparison between AI-heavy vs student-driven outputs Evaluation Instruments: - A mixed-method approach will be used: - Pre- and Post-Surveys: Measure changes in ethical awareness - Interaction Logs: Track usage patterns and decision shifts - Assessment Rubrics: Evaluate quality and originality of outputs - Reflection Responses / Think-Alouds:Capture student reasoning - Comparative Analysis:Control vs experimental group differences Conditions for Implementation: - Institutional Alignment: Integration with clearly defined AI policies - Controlled Environments: Pilot deployment in LMS or academic settings - Collaboration with Educators: To define acceptable AI use cases - Iterative Testing: Continuous refinement based on feedback A critical condition for success is how the system is perceived: It must function as a learning support tool, not a surveillance mechanism. If students perceive it as restrictive, they may disengage or bypass it. If positioned as supportive and transparent, it can foster trust, reflection, and responsible AI use. Key Insights: This approach shifts the focus from: “Did the student misuse AI?” to “How can AI guide better decisions while preserving learning?” Example: A student asks the AI: “Can you write a structured introduction that evaluates both positive and negative impacts of social media on adolescent mental health, with a clear thesis statement?” Without guidance (Control): The AI generates the introduction immediately → student copies it with little reflection. With guidance (Experimental): The AI responds: “Do you want help developing your ideas or should I write it for you? Your course allows the first.” The student then chooses how to proceed, becoming aware of ethical boundaries.