The 3-Axis Framework for Evaluating Human-AI Interaction: Exploration, Agency, and Understanding

Jun 21, 2026 | Articles, Artificial Intelligence, Education, Human Development

Author

Alexander M. Economides

Purpose

This article was written independently by Alexander M. Economides and originally published on LinkedIn. It was developed to provide a practical framework for evaluating human-AI interactions, shifting the discussion away from AI capabilities alone and toward the impact those interactions have on human learning, judgment, and intellectual development.

Click here to read the original article on Linkedin

Original Article

Artificial intelligence is already changing the way people work, learn, write, research, and create. Most discussions focus on the capabilities of AI systems: how intelligent they are, how accurate they are, or what jobs they may eventually replace. These are important questions, but they miss a more immediate one. How should we evaluate our interactions with AI?

I propose a simple framework based on three axes. The first axis, Exploration, describes what we are studying; are we using AI to explore the ideas of others, or are we using it to develop, challenge, and refine our own ideas? The second axis, Agency, describes who is directing the work; are we managing the tool, or is the tool managing parts of the process for us through planning, coordination, and guidance? The final and most important axis, Understanding, measures whether the interaction is moving us toward Mastery or toward Hubris. Mastery is genuine understanding; it is the ability to exercise judgment, recognize assumptions, identify failure modes, and ask increasingly better questions. Hubris is the illusion of understanding; it occurs when the user mistakes access to information, vocabulary, or answers for expertise.

The first two axes describe what is happening. The third axis describes whether the interaction is producing a desirable outcome.

This distinction matters because artificial intelligence dramatically lowers the cost of accessing and synthesizing information; that is a remarkable achievement. However, access to information is not the same thing as understanding. Expertise has always involved more than answers. It requires judgment, context, experience, and the ability to recognize what questions the user should still be asking.

Consider two students using AI throughout their semester. The first uses AI to complete homework assignments, generate answers, and improve grades. By the end of the course, the student has earned excellent marks on every assignment but fails the final in-person examination because they developed almost no genuine understanding of the subject. The second uses the same technology to challenge assumptions, identify weaknesses in reasoning, explore alternative explanations, and deepen their understanding. The first interaction moves toward Hubris. The second moves toward Mastery. And when the tool is taken away, the difference becomes visible. The technology is identical; the outcome is not.

The framework I propose is intentionally agnostic regarding the specific use case. It applies equally to students, researchers, consultants, programmers, executives, and policymakers. It does not attempt to determine whether AI is good or bad. Instead, it asks a simpler question: Is this interaction moving the user toward Mastery or toward Hubris?

That question may become increasingly important as AI systems improve. The greatest risk may not be that machines become more capable. Instead, it may be that humans become less capable while gaining false confidence in their capabilities.

A civilization can preserve vast amounts of information and still lose the ability to generate understanding. We should avoid the trap of evaluating the future of AI based solely on what the machines can do. We must also evaluate what happens to the humans who use them.

Framework Summary:

Axis A: Exploration (What is the user exploring?)

  • The Ideas of Others: The user uses AI to investigate external ideas, knowledge, frameworks, and existing thought.
  • Their Own Ideas: The user uses AI to stress-test, refine, and develop their own conjectures and frameworks.

Axis B: Agency (Who is directing the work?)

  • The User Manages the Tool: The user directs AI to produce outputs and artifacts.
  • The Tool Manages the User: The AI coordinates, sequences, organizes, and guides the user’s work.

Axis C: Understanding (What is happening to the user?)

  • Mastery: The user moves toward genuine understanding, judgment, and the ability to ask increasingly better questions.
  • Hubris: The user mistakes easy access to answers, vocabulary, or expertise with genuine understanding.

Every AI interaction occupies a position somewhere within this three-dimensional space. It is not enough to use AI effectively; we must also ensure that its use moves us toward Mastery rather than Hubris.

Author’s Note: This article is the second in a series examining the future relationship between artificial intelligence, knowledge, and human development.

Tags: Artificial Intelligence, Human Development, AI Ethics, Critical Thinking, Education, Mastery, AI Frameworks, Knowledge