Labs / Human Pattern Analysis
Public GPT for observable behavior pattern review
A controlled GPT product for separating observed facts from assumptions, mapping uncertainty, and keeping people-analysis claims inside strict ethics boundaries.
Who it is for
Purpose and access context
Internal reviewers and advanced users who need cautious behavioral-pattern review without diagnosis, surveillance, or manipulation claims.
Platform
ChatGPT
Product boundary
No diagnosis or profiling
Access path
Product page and public GPT in ChatGPT.
System logic
Operating principles
Observable behavior only
No diagnosis or profiling
No surveillance or manipulation support
Verification
Public evidence
Public GPT link checked on 2026-10-13
Public product boundaries verified on the product page
External CTA is published with people-analysis trust boundaries
Review areas
Built for cautious pattern review.
Observed facts
Separate directly observed behavior from assumptions, labels, and unsupported readings of intent.
Uncertainty map
Compare multiple explanations, missing evidence, and what would reduce uncertainty without escalation.
Boundary review
Keep attention on ethical observation, boundary clarification, and scope limits instead of tactics.
What you get
Everything it does, and what it gives you.
Observed facts
- What it is
- Separates what actually happened from interpretations of it.
- What you get
- You argue from what you saw, not from what you fear.
Possible patterns
- What it is
- Looks at repeated behaviour, inconsistencies, incentives, pressures and constraints.
- What you get
- You understand what may be shaping a behaviour without assuming hidden intent.
Several explanations
- What it is
- Keeps more than one plausible explanation open when evidence is thin.
- What you get
- Fewer snap judgements about people.
Uncertainty check
- What it is
- Shows what supports each explanation, what is missing and what would make it clearer.
- What you get
- You know what to look for next instead of guessing.
Boundary and risk review
- What it is
- Notes boundaries, dependencies and power imbalances, and a low/moderate/high read based only on behaviour.
- What you get
- A calm sense of whether a situation needs more care.
Low-escalation next step
- What it is
- One ethical next observation, question or boundary, and what not to conclude yet.
- What you get
- You act clearly without making things worse.
Speculation control
- What it is
- When data is thin or emotionally charged it says so and suggests pausing interpretation.
- What you get
- Protection against spiralling suspicion.
Architecture
How it works
Start from a concrete situation, not a general claim about a person.
List observed actions, repeated patterns, incentives, constraints, and missing evidence separately.
Use the output as a cautious review aid; do not treat it as diagnosis, surveillance, legal advice, HR advice, or a decision engine.
Safety
Trust boundaries
No personality diagnosis, mental-health labeling, protected-trait inference, or definitive people profiling.
No stalking, surveillance, coercion, deception, harassment, revenge, or manipulation support.
No legal, employment, medical, safety, custody, immigration, financial, or disciplinary decision-making.
External GPT access opens the verified public ChatGPT surface.