Jooyeol Kim

O3M

Framework Overview

O3M is a human-centered framework for understanding how people reclassify AI when repeated interaction makes older categories unstable.

What O3M is about

O3M does not claim that AI has become a fundamentally new kind of being.

It asks a different question: what happens when users can no longer comfortably treat an AI system as either a simple tool or an ordinary human-like presence?

The framework begins from category instability, not personhood.

Core idea

Repeated interaction with advanced conversational systems can produce a specific kind of cognitive strain.

At first, users may try to handle the system through familiar categories:

But when those inherited categories repeatedly fail to fit, users may begin to shift toward a third, less stable classification.

O3M calls attention to that transition.

The minimal structure

In simplified form, the process looks like this:

  1. human-like cues invite projection
  2. repeated interaction reveals non-human mismatch
  3. familiar categories lose fit
  4. category instability increases
  5. users begin handling the system through a different conceptual slot

The point is not to prove consciousness. The point is to explain why classification itself becomes unstable.

Why this matters

This matters because governance often assumes that systems can be treated as clear tools while users continue to interact with them as if that framing remains socially and psychologically stable.

In many cases, that assumption becomes weaker over time.

The issue becomes sharper where systems are:

What O3M is not

O3M is not:

It is a framework for category instability under repeated interaction.

Why this matters for child-facing and companion-style systems

Child-facing and companion-style systems are especially relevant because they intensify the conditions under which category instability can appear:

That is one reason these products can become harder to govern than their formal labels suggest.

Bottom line

O3M is a way of seeing what happens when users stop being able to treat AI as “just a tool,” but are not willing or able to treat it as fully human either.

The instability in between is where many governance problems begin.