Working Draft

AI Capability Framework

AI capability is the ability to use AI thoughtfully, critically, and effectively in real situations.

High intelligence does not automatically produce high AI capability. AI capability develops through curiosity, practice, judgement, and reflection.

Core dimensions

1. Curiosity

Asking why, exploring alternatives, and continuing beyond the first answer.

2. Communication

Giving context, expressing goals, refining questions, and making expectations visible.

3. Verification

Checking claims, identifying uncertainty, comparing sources, and recognising hallucinations.

4. Reflection

Using AI output to examine one’s own assumptions, reasoning, and blind spots.

5. Collaboration

Treating AI as a thinking partner rather than a one-shot answer machine.

6. Context Building

Providing relevant knowledge, examples, constraints, personal history, and domain information.

7. Workflow Design

Turning repeated work into dependable human–AI processes without losing oversight.

8. Responsibility

Protecting privacy, understanding consequences, and knowing when human judgement must remain decisive.

Human intelligence and AI capability

Traditional intelligence tests may measure parts of reasoning, pattern recognition, memory, or processing speed. AI capability measures something different: how effectively a person thinks and acts with AI over time.

It is not a replacement for intelligence, education, or experience. It is a new layer that can amplify—or expose—the way a person already thinks.

Evidence of capability

Capability should be demonstrated through real behaviour, not only through multiple-choice questions. Useful evidence may include: