top of page

AI Augmentation Roles: Choosing How AI Supports Thinking, Learning, and Decision-making

Oct 29
5 min read

Updated: Oct 31

Using AI effectively isn't just about knowing how to write a good prompt. It's about knowing what role you want AI to play in supporting your thinking, learning, and decision-making.


Too often, we approach generative AI in the same way: ask a question, receive an answer, and move on. But AI can offer much more than information or completed tasks. It can challenge assumptions, guide reasoning, develop skills, support creativity, or help us work within increasingly AI-enabled environments.


The important question is not simply "What can AI do for me?" but "How should AI support me, and what capabilities am I trying to develop?"


This is where understanding AI augmentation roles becomes valuable. AI augmentation refers to using artificial intelligence to support, extend, and enhance human capabilities, rather than replace human thinking, judgment, and decision-making.


Connecting AI Augmentation to CAC Co-intelligence


Consider the 3-stage CAC Co-intelligence process:


Cognitive → AI → Cognitive, or more simply, Thinking → AI → Thinking


Users first engage independently with a task, developing their initial understanding, reasoning, or ideas. In Stage 2, AI is introduced purposefully to support, extend, or challenge that thinking. Finally, they return to independent evaluation, synthesis, and reflection.


CAC Co-intelligence
CAC Co-intelligence

Simply acknowledging the CAC process is not enough. Users also need to understand how AI can support their cognition during Stage 2. Different AI augmentation roles provide different ways of supporting and extending human thinking, learning, and decision-making. Each approach shapes how we engage with AI, the contributions AI makes, and the cognitive responsibilities we retain. Understanding these differences allows us to make more purposeful choices about how AI can support our activities, while ensuring its use strengthens, rather than undermines, the knowledge and capabilities we are trying to develop.


Six AI Augmentation Roles


The following six augmentation roles introduce different ways AI can support learning and professional practice. For those new to AI, these examples provide a starting point for understanding the different approaches available, how each changes the way we interact with AI, and the benefits and limitations of each. By exploring when and why to use these roles, we can begin to make more informed and purposeful decisions about how AI can best support our thinking, learning, and work.


Six AI Augmentation Roles
Six AI Augmentation Roles

A note about the example prompts: These are deliberately simple, introductory prompts designed to illustrate the fundamental purpose of each augmentation role. They are starting points, not comprehensive prompting strategies. More sophisticated applications benefit from carefully developed prompts that establish relevant context, intended learning outcomes, the learner's existing understanding, appropriate boundaries, and expectations for AI interaction. The quality of the interaction depends on how effectively the prompt supports the intended cognitive activity, not simply its length.


1. Socratic AI — Deepening Understanding Through Questioning


Rather than supplying answers, AI asks purposeful questions that encourage learners to explain their reasoning, examine assumptions, and identify gaps in understanding.


Why use it? To develop reasoning, metacognition, and deeper conceptual understanding while keeping students cognitively engaged.


When to avoid it: When students lack essential foundational knowledge and require explanations before they can meaningfully engage with questions.


Example prompt:


"I'm learning about correlation and causation. I've studied the concepts, but I'm struggling to understand the difference. Ask me one question at a time to help me examine my understanding. Build on my responses, challenge my reasoning, and don't simply give me the answers."





2. Centaur — Strategically Dividing Work Between Human and AI


Centaur focuses on the deliberate division of labour between humans and AI.


Why use it? To develop strategic judgement about what to delegate, what to undertake independently, and where collaboration adds value.


When to avoid it: When delegating a task would bypass the thinking or skill the student needs to develop.


Example prompt:


"I've identified the tasks required to complete my project and developed an initial plan for dividing them between myself and AI. Help me evaluate this division based on our respective strengths, limitations, and risks. Suggest alternatives, but leave the final decisions to me."




3. Adversarial AI — Strengthening Thinking Through Challenge


AI acts as a constructive opponent, challenging arguments, questioning evidence, and presenting credible counterarguments.


Why use it? To strengthen critical thinking, expose weaknesses, test assumptions, and develop more defensible judgements.


When to avoid it: When learners are still establishing basic understanding or when excessive criticism could create confusion rather than productive reflection.


Example prompt:


"Act as a sceptical senior manager reviewing my recommendation to redistribute staff responsibilities. Challenge my assumptions, question the strength of my evidence, and present credible counterarguments. Allow me to defend or revise my position before offering further criticism."




4. AI Tutor — Developing Knowledge and Skills


AI takes on a teaching role, providing explanations, scaffolding, guided practice, feedback, and progressively challenging activities.


Why use it? To support personalised learning, address misconceptions, and progressively develop knowledge and skills. AI Tutors may combine several augmentation techniques, including Socratic questioning and constructive challenges.


When to avoid it: When excessive guidance risks creating dependence or when students must demonstrate their abilities independently.


Example prompt:


"Act as my personal tutor. I'm learning statistical hypothesis testing but struggling with p-values. Start by checking my understanding, then provide simple explanations, guided practice, and feedback. Adapt your support as I improve, but don't give me answers before I've attempted the questions myself."


Note: A detailed tested tutor prompt can be found here.




5. AI Environment — Working Within AI-Enabled Systems


AI capabilities, including agents, are embedded within learning platforms, professional tools, and interconnected workflows.


Why use it? To develop the ability to work effectively with embedded and agentic AI, understand how these systems operate, evaluate their contributions, and maintain informed oversight.


When to avoid it: When systems lack transparency, appropriate safeguards, or opportunities for meaningful human control, particularly in high-stakes activities.


Example instruction for an embedded AI agent:


"Monitor project milestones and flag potential delays. Identify affected tasks, explain their likely impact, and suggest possible adjustments. Use only the project information I've authorised, and don't make changes without my approval."


This example assumes an AI-enabled project management system with access to authorised project information and configurable agent capabilities.




6. Co-Creation — Developing Ideas Together


Human and AI work iteratively to generate, combine, and refine ideas or solutions.


Why use it? To support creativity, synthesis, innovation, and the exploration of alternative possibilities.


When to avoid it: When students need to demonstrate independent creativity or when AI-generated ideas might prematurely constrain their thinking.


Example prompt:


"I have designed and uploaded a basic idea for SOMETHING. Act as my creative partner. Suggest three ways to develop my concept, then help me generate and refine a design based on my feedback. Let me guide the creative decisions."




Choosing the Right Augmentation Role Matters


These augmentation roles are not mutually exclusive. An AI Tutor might incorporate Socratic questioning, while an AI-enabled professional environment might combine several augmentation approaches within everyday workflows.


The key is to start with your purpose, not the AI capability.


What are you trying to achieve? Which activities would benefit from AI support? What knowledge, skills, or professional judgement must remain your responsibility? And how might different approaches to AI engagement influence the quality of your thinking, learning, or work?


Within education, CAC Co-intelligence provides an overarching cognitive process for ensuring AI supports rather than replaces learning. The augmentation roles help us make more deliberate choices about how AI contributes within that process. These same principles extend to professional practice, where selecting an appropriate AI role can strengthen problem-solving, creativity, decision-making, and the effective use of AI-enabled workflows.


Ultimately, effective AI augmentation isn't simply about achieving better outcomes with AI. It's about developing and maintaining the human knowledge, skills, and judgement needed to work effectively alongside it.


New to the 3-stage CAC Co-intelligence process? Click the link to find out more.


A/Prof Sasha Nikolic

University of Wollongong

29/10/26


Each AI Augmentation role can support you in different ways
Each AI Augmentation role can support you in different ways

Comments


aaeeHorizontal.jpg

This is a Special Interest Group of the Australasian Association of Engineering Education

https://aaee.net.au/

Follow Us

  • LinkedIn

© 2024 The Australasian Artificial Intelligence in Engineering Education Centre. All Rights Reserved.

bottom of page