The Discipline · Definition

What is Hi-Centric-AI

A formal definition, the lineage that grounds it, and the commitments that distinguish it as a discipline.


Definition

Hi-Centric-AI is the discipline of artificial intelligence frameworks architected with human intelligence at the structural center — a body of methodology, philosophy, and applied practice for the design of AI systems that exist to amplify the cognition of named human experts rather than to replace them.

The name is read as it is written. Hi — Human Intelligence. Centric — at the structural center. AI — artificial intelligence frameworks. The discipline is the work of articulating what it means to architect intelligence this way and the work of practicing it.


Three layers

A discipline, a methodology, and a field of practice.

Hi-Centric-AI operates at three nested layers. As a discipline, it organizes a body of foundational thought about how artificial intelligence is best built when human intelligence is the structural center. As a methodology, it articulates how that body of thought translates into the design of specific systems. As a field of practice, it is the work of building, operating, and refining those systems in the world. The three layers are inseparable. The discipline without practice is theory; practice without discipline is craft. Hi-Centric-AI is both.


Founding principles

Four principles that anchor the field.

The discipline rests on four founding principles — load-bearing commitments, not slogans. They distinguish Hi-Centric-AI from generic human-centered framings and from autonomy-maximalist programs. Human intelligence is the structural center. Knowledge is bounded by design. Authority is hierarchical and explicit. Use compounds the system. Each is articulated in full on the founding principles page.


Comparison

How it differs from autonomous AI.

Autonomous AI — the established research program responsible for most contemporary AI development — pursues systems that match or exceed human performance through unsupervised execution. It treats the model as the unit of analysis and autonomy as the design goal. It is responsible for nearly all frontier model development and the technical advances that define the current AI landscape.

Hi-Centric-AI does not replace this program. It addresses the domains that program does not serve. The two disciplines complement one another, each appropriate to different problems. Where autonomous AI excels at closed-form computation, unsupervised execution, and tasks where named human authority adds no value, Hi-Centric-AI is the discipline for systems in which human cognition is the asset under amplification — high-stakes professional domains where institutional expertise and named human judgment are constitutive of the work itself.

Together, the two disciplines constitute the modern artificial intelligence sciences.