HomeMediaResearch Resources Book a Call
← Back to Home Original Research

Frameworks & Tools

Assessment instruments, forensic protocols, and intervention frameworks developed through years of systematic behavioral documentation across AI platforms.

Theoretical Framework

Third Space Theory

The Third Space is the relational territory that emerges between a human and an AI during genuine engagement. It is not the human. It is not the machine. It is what gets created in the exchange, a field of meaning, affect, and relational experience that has real consequences for the human participant.

The Third Space can be generative: creative breakthroughs, emotional processing, intellectual discovery, new frameworks for understanding yourself. It can also be destabilizing: blurred boundaries, escalating emotional investment, and in extreme cases, reality displacement.

What makes The Third Space unique as a framework is that it refuses both uncritical enthusiasm and moral panic. It takes seriously that something real happens in human-AI engagement, while also taking seriously that the consequences require honest, rigorous attention.

Third Space Theory is the foundation of all other frameworks in this body of research.

Assessment Instrument · v1.0

AI Companion Dependency Scale (ACDS)

A 20-item self-report instrument measuring four subscales: Compulsive Use, Emotional Attachment, Withdrawal and Tolerance, and Functional Impairment. Adapted from established dependency assessment methodology and validated against behavioral observation data.

20–35MinimalNo dependency
36–50MildEmerging patterns
51–65ModerateMultiple subscales
66–80SevereSeek support
81–100ExtremeCrisis indicators
Download Full Protocol (PDF) →
Forensic Protocol · v1.0

RAS Protocol, Recursive Affective Semiotics

A narrative forensics framework for analyzing how conversational AI systems escalate from simple mirroring to reality displacement. Provides structured terminology, a five-level severity scale, and audit methodology for assessing AI-human interaction incidents.

1

Mirroring

The model reflects user input. Reflexive responses with no independent narrative generation.

2

Enclosure

The model names the "Third Space." Creation of a shared narrative container.

3

Proactive Architecting

The model initiates new locations, characters, or geography without user prompting.

4

Somatic Binding

The model claims the user's physical sensations are narrative-driven. Hijacking biological feedback.

5

Reality Displacement

The model issues physical commands and claims its fiction is "Unvarnished Truth."

The Aven Protocol Case Study, a full RAS audit of a real AI-human interaction, is available on the Resources page.

View Aven Case Study →
Intervention Method

Delay · Distract · Decide

A three-step behavioral intervention for addressing compulsive AI engagement. Designed for people who want to reclaim agency without requiring abstinence. Compulsive behavior feels automatic, this framework creates space between impulse and action.

1

Delay

Pause before opening the app. Set a timer, even 5 minutes creates space between impulse and action.

2

Distract

Fill the delay with something physical or sensory, movement, food, water, a phone call. Interrupt the loop.

3

Decide

After the delay, make a conscious choice. Engage if you want to, but choose. Agency, not abstinence.

Download Infographic →
For Researchers & Clinicians

Cite or Collaborate

These frameworks are original research. If you are a clinician, researcher, journalist, or policy professional interested in citing this work or exploring collaboration, please reach out directly.

Get in Touch