CAC Co-intelligence:
Learning with AI, Not from it
Case Studies
From CAC Theory to Teaching Practice
On the previous pages, you have explored why CAC Co-intelligence is needed and how the framework works, including how students can progressively learn to use it. The next challenge is implementation.
In my experience, this is often where educators need the most support. Understanding the principles of CAC is one thing; visualising what a CAC learning experience might actually look like in your own classroom is another. Questions quickly emerge:
-
What should students know before they begin?
-
What thinking should happen before AI is introduced?
-
What role should AI play?
-
How do I stop AI from replacing the learning?
-
How should students evaluate its contribution?
-
And how do I know that the student, not the AI, has actually learned?
To help make this more concrete, I have developed a collection of fictional, discipline-specific CAC case studies. The examples were developed with the assistance of generative AI and then structured around the CAC Co-intelligence framework. They are designed to make the learning-design process visible: from establishing the necessary disciplinary foundations, through the first Cognitive stage, into a deliberately chosen AI augmentation role, and then back to human evaluation, refinement, validation and responsibility.
The case studies are not intended as model lessons that should simply be copied into a subject. Instead, think of them as worked design examples. Each one demonstrates the kinds of decisions an educator needs to make when designing a CAC experience: what cognition must be protected, what AI can usefully augment, which risks and ethical implications matter in that context, what evidence students should use to validate AI output, and how independent learning can ultimately be verified.
The examples span a range of disciplines including mathematics, science, commerce, law, education, engineering, computer science and the arts, They demonstrate that CAC is not tied to one type of task or one type of AI use. In different cases, AI may operate as a Socratic tutor, critic, alternative-method generator, adversarial reviewer, misconception detector, source critic or design reviewer. What remains consistent is the underlying principle: AI should support, extend or challenge human thinking without unnecessarily replacing the cognition required for learning.
The diagram above provides a high-level view of the decisions embedded within a CAC learning experience. The downloadable case studies below show what those decisions can look like when translated into actual learning activities. As you explore them, focus less on the particular disciplinary content and more on the design logic: what happens before AI, why AI is introduced, the role it is given, how its contribution is evaluated, and how responsibility returns to the learner.
GenAI Created Case Studies
Engineering:
The statics case uses AI to question free-body diagrams and equilibrium reasoning, while the design case uses AI to challenge material-selection trade-offs. Both emphasise verification, safety and accountable engineering judgement
Commerce:
The accounting case uses AI to challenge accrual and transaction reasoning, while the economics case uses AI to critique a pricing decision and its assumptions. Both demonstrate how AI can broaden analysis without taking over the final judgement.
Mathematics:
Two first-year examples show CAC in action: calculus, where AI acts as a Socratic tutor to deepen conceptual understanding, and linear algebra, where AI critiques and compares alternative solution methods. Both protect mathematical reasoning while developing evaluative judgement.
Education:
One case focuses on interpreting student thinking, with AI challenging premature assumptions. The second examines formative assessment design, using AI to test alignment, fairness and accessibility. Both emphasise professional judgement, privacy and responsible use.
Computer Science:
The debugging case uses AI to challenge hypotheses and test design rather than fix code. The algorithms case uses AI to critique correctness, complexity, memory use and alternatives. Both protect core computing skills while using AI as a reviewer.
Law:
The first case uses AI to challenge case analysis and precedent, while the second uses AI as an adversarial critic in statutory interpretation. Both reinforce that AI may support legal reasoning, but legal authority must come from recognised sources.
Science:
The biology case uses AI to question explanations of enzyme activity, while the chemistry case uses AI to critique an experimental titration method. Together they show how AI can support scientific reasoning while evidence and disciplinary validation remain central.
Arts:
The literary studies case uses AI to challenge close reading and alternative interpretations, while the history case uses AI to test sourcing, context and corroboration. Both show how AI can broaden perspective without replacing evidence-based interpretation.
Look for the design decisions, not just the activity:
The most important part of each case study is not the particular question or AI prompt. It is the reasoning behind when AI is introduced, what role it is allowed to play, what remains human, and how learning is subsequently verified.
Create your own case study
The examples are intended to help you see what is possible, but CAC needs to be designed around your learning objectives, your disciplinary practices, your students and your context. To help you begin that process, I have also developed a reusable prompt that can be copied into any generative AI platform.
You only need to provide a small amount of information about your degree, subject, year level and topic. The prompt then guides the AI through the same design considerations used in the case studies, including the prerequisite teaching students need, the cognition that should occur before AI, an appropriate augmentation role, ethical considerations, evaluation and validation, cognitive risks, and independent learning verification.
The generated output should be treated as a starting point for professional judgment and not a finished learning design. Review it against your disciplinary expertise, curriculum requirements, institutional policies, student needs and assessment context before implementation. In other words, the prompt itself should be used in the spirit of CAC: think first, use AI purposefully, then evaluate and refine what it produces.
Real Case Studies
This is where I ask the community for help. If you use CAC Co-intelligence please share what you did and how you did it, by sending me an email to get the conversation started. The community needs your help, I need your help.
​
Over time, I will also be uploading my own case studies to the website.