In a special article for World ORT, former Hatter Technology Seminar Alumna, Pedagogical Leader and IBO Certified Educator, Victoria Guevara, examines the importance of effective AI-resilient instruction
After 18 years of teaching experience, I’ve learned that my convictions are never final. I have radically changed my mind and my practice shaped by emerging technologies, hours and hours of professional learning, and most powerfully by actual students in actual classrooms day in and day out.
While the world outside the classroom operates between a constant process of fracture and repair, every learner, every cohort, every class period seems to carry its own ecosystem of attention, resistance, curiosity, and silence. So what follows are simply ideas I hold about teaching and learning today.
In a world saturated by artificial intelligence (AI), critical thinking is a survival skill. While AI systems can generate answers, summarize texts, and simulate reasoning, they do not independently decide what deserves trust, what carries ethical weight, or what meaning signifies for a particular human life. The instruction that matters for the AI generation must therefore shift from asking students what they know to examining how they know, why they trust certain sources, and on what grounds they are willing to stand behind a claim.
Effective AI-resilient instruction, the one that survives the convenience of automation, emphasizes inquiry-based learning, thinking routines, and metacognitive awareness. Students must be taught explicitly and repeatedly to analyze, interpret, evaluate, infer, synthesize and justify ideas using predetermined thinking taxonomies. These skills are not assumed to develop organically; they are modeled, practiced, and reflected upon across disciplines. Reflection checkpoints such as process journals, portfolios, and staged drafting and outlining make progress traceable and support students to take ownership of their intellectual growth and unique perspectives.

Reflection, more than an end-of-unit ritual, must be taught and consistently modeled as a discipline of mind rather than treated as a personality trait few students possess. As Professor Jenelle Kirchoff argues in her work on helping students “reflect, record, and own” AI-assisted learning, structured reflection is essential for making thinking visible and protecting authentic authorship in environments in which technological influence can otherwise remain invisible (Kirchoff ¹, 2025).
What this looks like in practice is not vague encouragement but sustained engagement with concrete questions: What problem am I actually trying to solve? Which ideas are mine and which came from AI? How did I verify the accuracy of this information? Did AI overlook a key perspective? What choices did I make to reshape this into my own thinking? When students are guided to articulate statements such as “AI helped me brainstorm, but I changed the suggestion by…” or “The final product is mine because…”, they are not merely complying with an academic honesty requirement; they are rehearsing the cognitive habits of authorship, judgment, and self-awareness.
Kirchoff’s emphasis is on embedding reflection checkpoints across the entire process, from inquiry to drafting to final authorship. That way, reflection grows into the mechanism through which learning becomes conscious, ethical, and transferable. In classrooms where teachers model this same vulnerability and transparency about their own thinking, reflection would then be less about surveillance and more about ownership and a disposition to reclaim one’s voice.
Additionally, a critical dimension of a rich teaching-learning process is source analysis. In an environment in which AI-generated content can appear authoritative, students must learn to apply frameworks to interrogate credibility, evaluate accuracy, relevance, bias, and reliability as instinctively as they once learned to memorize definitions.

Comparing AI outputs with teacher-curated resources, trusted databases, and prior knowledge strengthens discernment and reinforces the principle that AI should support thinking, not replace it. Ultimately, the goal is to develop learners who take pride in their reasoning rather than in polished, externally generated answers.
This capacity to critically interact with content, however, does not flourish in isolation. It requires coherence between goals, assessments, and learning experiences. Frameworks such as Understanding by Design insist that teachers begin not with activities, but with meaning: What should students understand deeply? What should they be able to transfer beyond this unit, this subject, this year? From there, we design tasks that require students to apply understanding in contexts that feel real, immediate, and consequential to their lives, families and communities.
Authentic assessments move beyond superficial performance. They ask students to make decisions, to argue, to reflect, to solve messy problems with no single correct answer. They require personal context, ethical judgment, intellectual risk and empathy.
Consequently, we must strategically guide students to disclose how they used AI in the process, to name the tool, explain the choice and evaluate its impact. We urge the classroom culture to begin to shift so transparency replaces suspicion, and academic integrity becomes a shared responsibility, a collective need.
And, collectively, we must acknowledge that at the heart of deep learning lies conceptual understanding. Concept-based teaching reframes curriculum not as a sequence of topics to be covered, but as a structure of ideas to be understood. Rather than treating facts and skills as endpoints, this approach organizes them through concepts, generalizations, and principles that allow learners to perceive patterns and construct meaning.
The difference is subtle but profound. When facts are readily available, conceptual understanding acts as a cognitive lens that allows students to organize information, connect ideas, and see relevance across disciplines and contexts. Facts remain essential, but they serve as evidence and support for conceptual understanding rather than as endpoints of learning.

A central mechanism in this process is what Erickson² (2009) describes as synergistic thinking: the deliberate interaction between factual knowledge and conceptual understanding. Students are not merely asked to define ideas, but to use them, to test them, apply them, interrogate them across time, culture and situation. This kind of thinking builds durable mental structures that support both retention and flexibility.
However, the clearest indicator that this structure exists is transfer. In the context of AI, transfer must be the protagonist. Facts tend to remain bound to the contexts in which they were learned. Concepts, by contrast, travel. When students grasp ideas at a conceptual level, they can apply them to new technologies, unfamiliar disciplines, and unpredictable real-world challenges. In a world in which new problems emerge faster than new curricula, this adaptability is not optional. It is foundational.
Transfer also operates across cultures. Concepts such as power, identity, sustainability, and conflict may appear differently in different contexts, yet their underlying structures persist. When students learn to recognize these patterns, they become capable of engaging thoughtfully with perspectives beyond their own. They develop not just academic competence, but a consciousness with which to foster empathy, connection and belonging.
For me, an AI-resilient education rests on the development of strong thinking skills that enable people to ask important questions, engage in deliberate reflection, conduct rigorous source analysis, and apply learning through transfer. Critical and creative thinking allow students to challenge assumptions and generate solutions. Information literacy equips them to navigate digital ecosystems ethically and intelligently. Reflection cultivates metacognition, which is essential for growth, while the capacity to transfer comprehension bridges learning across subjects and contexts.
Overall, we want learners who develop agency, who begin to see themselves not as recipients of knowledge, but as active creators of meaning. This leads us, inevitably, to the role of the teacher in the classroom. Education for the AI generation demands careful learning design. We must build environments that privilege depth over coverage, inquiry over compliance, and deep reflection over our obsession with performance. In such a model, AI is neither banned nor glorified. It is a tool: useful for modeling, drafting, comparing, provoking, refining. Not only do students learn how to use it, but also when to distrust it, when to challenge it, when to step away from it. Ideally, this produces learners who engage with technology with genuine curiosity and intellectual integrity.
Ultimately, the goal is to develop what AI cannot yet replicate: human judgment, conceptual insight, ethical reasoning, and the ability to transfer understanding across the unpredictable challenges of the future.
Work Cited
¹ Kirchoff, J. (2025, October 14). Whose words? Teaching students to reflect, record & own AI-assisted work [Presentation]. International Baccalaureate Organization.
² Erickson, H. L., & International Baccalaureate Organization. (2012). Concept-based teaching and learning [IB position paper].











