Rye Country Day School
Making responsible progress toward AI proficiency without sacrificing scholarship
The AI & Cognitive Learning Framework reinforces the same four capacities of the Portrait of a Graduate: the Critical Thinker who questions AI outputs and sources, the Reflective Collaborator who treats AI as a partner rather than a ghostwriter, the Authentic Communicator whose voice drives the final results, and the Compassionate Citizen who weighs the ethical and environmental implications of using these tools. The framework gives teachers and students a shared language for protecting those capacities as AI becomes more integrated in academic work.
An active mental process focused on how thinking, evaluating, and problem solving connect new information to existing knowledge.
Actively fact-checking, assessing sources, claims, and ideas for accuracy, bias, and quality.
Never sharing personal information, being transparent about the use of AI tools, and understanding its environmental impact.
Working with peers, teachers, and experts to deepen understanding and explore new opportunities to solve problems.
Understanding how AI works, where it succeeds, and where it falls short.
All thinking and submitted work is fully independent of AI.
In-class assessments, tests, in-class essays, timed writing. Default for any unspecified assignment.
Over-reliance on recall without understanding; difficulty organizing independent thought when AI scaffolding is absent.
AI is used only to explain concepts or define terms. It should not generate ideas, brainstorm, or outlines.
Homework reading, studying for tests, concept review. AI acts like a dictionary or tutor.
Watch for scope creep: students sliding from "explain this concept" into "explain what I should argue." The teacher should speak with the student when this happens.
AI may assist students in pushing their own thinking further and building on ideas they have already developed independently.
Low-stakes homework assignments, shorter papers, and problem sets. Student submits an initial idea, concept, or outline, then uses AI to challenge it.
Copying AI-generated rewrites verbatim; losing student voice; AI changes that subtly alter meaning without the student noticing.
AI offers feedback, new ideas, and possible gaps on a draft the student has already completed.
Long-term projects, research papers, and data collection for presentations. Student completes the assignment, then uses AI for structured feedback.
Accepting AI ideas without reviewing/vetting them; AI shaping the student's thinking rather than the student directing AI.
AI generates content that the student curates, evaluates, and synthesizes into a disclosed final product.
Teacher-unlocked only. Complex projects, creative and media work, or professional workflow simulation. Full disclosure is mandatory.
Students relinquishing the core learning challenge to AI. Productive struggle disappears when AI becomes the author. Students must stay in the director's seat.
Level One
All thinking and submitted work is fully independent of AI.
No-AI tasks are where this competency shows up most clearly. Students have to generate, organize, and sustain their own thinking with no external help. That difficulty is not a flaw in the design; it is the design. Durable understanding develops through the work of thinking, not around it.
When students produce work entirely on their own, they also have to judge it on their own. That means evaluating the quality of their own arguments and reasoning without anyone or anything to lean on. That is a harder skill than it sounds, and it only develops through practice.
Establishes authentic authorship. Students need regular practice producing work on their own to develop real ownership of their ideas, arguments, and voice.
Builds cognitive stamina. Organizing a complex argument, working under pressure, staying focused through difficulty: these are things students get better at by doing them, not by avoiding them.
Enables valid assessment. Teachers can only know what a student understands if the student produces it unassisted.
Honors the learning process itself, not just the product.
Level Two
AI is used only to explain concepts or define terms. It should not generate ideas, brainstorm, or outlines.
Keeping AI strictly to clarification keeps the student in the role of thinker. When students have to show they understood something in their own words after consulting AI, the clarification becomes a support rather than a shortcut.
Getting an AI explanation is only useful if the student actually checks it. Is it accurate? Does it actually answer the question? Holding AI output up against their own understanding is a basic but important habit, and this is where students start building it.
Even at the clarification level, students must be transparent that AI was consulted and avoid sharing personal or sensitive information in prompts. Naming the tool and the question keeps the support honest.
Supports self-directed learning without replacing thinking. Asking AI to define a term is like using a dictionary: a reference, not a ghostwriter.
Teaches productive AI literacy. Formulating a precise clarifying question is itself a skill.
Democratizes access to explanation for students without strong home tutoring support.
Keeps the student as the thinker. AI answers the question the student poses; it does not generate the thesis or structure the argument.
Level Three
AI may assist students in pushing their own thinking further and building on ideas they have already developed independently.
The human-first draft is the anchor. Students must evaluate AI suggestions against their own prior thinking and decide what to keep, modify, or reject. That is where real learning happens.
AI becomes a thinking partner, but only if the student treats it as one. That means pushing back, asking follow-ups, and staying critical. Collaboration is not agreement.
Students begin to see how AI generates ideas, where it defaults to generic advice, and how framing a question differently changes the output. This is practical AI literacy.
Students must disclose AI use, protect personal information, and make sure AI suggestions never quietly override their own authorship or voice.
Ensures AI strengthens the student's thinking rather than replacing it. The human-first draft is a non-negotiable safeguard.
Develops metacognitive skill. Students must reflect on how AI shaped their thinking and justify their final direction.
Models real-world professional practice. Many professional environments use AI this way: brainstorm first, then stress-test with AI.
Prevents AI dependency. By requiring independent thinking before AI engagement, students build and maintain their own intellectual capacity.
Level Four
AI offers feedback, new ideas, and possible gaps on a draft the student has already completed.
AI feedback is not automatically correct. Students must evaluate whether AI suggestions actually improve their work or dilute their voice and argument. That judgment is the skill.
Using AI for feedback requires discipline. The line between "AI suggested a revision" and "AI rewrote my paragraph" is real, and students must stay on the right side of it.
Treating AI as a collaborator means engaging with its feedback critically, not accepting it wholesale. The student remains the author and the decision-maker.
Students develop a more nuanced sense of AI's strengths and limitations when they see how it responds to finished work. They learn where AI feedback is useful and where it falls flat.
The completed draft proves the student can produce independent work. AI feedback comes after, not during, the creative process.
Annotation and reflection requirements make the revision process visible and accountable.
Teaches students to be critical consumers of feedback from any source, not just AI.
Models a professional revision workflow where peer review and AI review can coexist with clear attribution.
Level Five
AI generates content that the student curates, evaluates, and synthesizes into a disclosed final product.
At this level, evaluation is everything. Students must assess AI output for accuracy, relevance, bias, and quality, then make deliberate choices about what to keep, modify, or discard.
Full disclosure is non-negotiable. Students must document exactly what AI generated, what they changed, and what they authored independently. This is both an ethical and an intellectual requirement.
The student remains the director. AI generates; the student curates. That distinction must be maintained throughout the project, and the authorship checkpoints exist to ensure it.
This is where AI literacy becomes most practical. Students must understand what AI can and cannot do, manage its outputs effectively, and reflect on how it shaped the final product.
Teacher-unlocked requirement ensures this level is only used when the learning goals genuinely benefit from co-creation.
Disclosure and documentation requirements prevent AI from becoming invisible. Everything is transparent.
Prepares students for professional environments where AI co-creation is increasingly common, but always with accountability.
The authorship checkpoint system ensures students maintain intellectual ownership even when using AI-generated material.