The coupled system as the unit of analysis

Combined Intelligence

A model for studying task-bounded human–AI systems in which outputs from either participant change the other participant’s next representation, judgment, or action.

Status and boundary: Descriptive theoretical framework. The book explicitly makes no claim to have coined the phrase and calls for comparative testing.

Architecture

How the framework is organized

1

Couple Around a Task

Define a bounded epistemic, creative, or practical task rather than attributing a global trait to the person–AI pair.

2

Create Reciprocal Interaction

Require iterative exchange in which both human and machine contributions alter what happens next.

3

Preserve Orientation

Keep purpose, evidence ranking, boundaries, consequence, and accountability explicit rather than letting fluent generation determine direction.

4

Compare Baselines

Evaluate the coupled system against relevant human-alone, AI-alone, or alternate-system baselines instead of assuming that a good artifact proves synergy.

5

Separate Capability from Wisdom

Higher performance does not establish moral maturity, truth, consciousness, or human worth.

Applications

Where the framework can support thinking

Human–AI researchOrganizational AI designEducation and learningProfessional augmentationCollective intelligenceGovernance and ethics

What it does not do

The framework is presented at the level supported by the source manuscript. It does not become a diagnostic test, a guarantee, a professional license, or an automated decision authority merely because it has a memorable structure.

Use it to improve inquiry, make assumptions visible, and organize human judgment. High-stakes legal, clinical, employment, financial, safety, religious, or regulated decisions still require the appropriate qualified authority.