19/01/2026
MAGG™
A Human-First Reactive Safeguarding Framework
Ethics, Constraints, and Non-Coercive System Design
Author: L-T-J-Dimmer - Me,Myself & BPD
Date: January 2026
Status: Draft (Framework Complete, Expansion Optional)
Abstract
This thesis presents MAGG™, a human-first safeguarding framework designed to prevent harm before escalation without diagnosis, surveillance, or authority over individuals. MAGG introduces Reactive Safeguarding as a bounded, time-limited stabilisation mechanism that responds to situational risk rather than personal interpretation. The framework prioritises dignity, freedom, reversibility, and non-coercion, embedding ethical constraints directly into system behaviour rather than relying on policy or user compliance. MAGG is positioned as an in-between interface — neither passive policy nor controlling intervention — offering a defensible alternative to extractive, diagnostic, and optimisation-driven systems.
1. Introduction
Contemporary safeguarding systems frequently rely on interpretation, categorisation, behavioural analysis, or surveillance to manage risk. While often well-intentioned, these approaches introduce secondary harms including stigma, authority drift, loss of agency, and coercive dependence.
This thesis proposes MAGG™ as an alternative design paradigm: a framework that stabilises risk without analysing people, extracting data, or enforcing behavioural correction. MAGG is not a therapeutic, diagnostic, or instructional system. It is a protective interface that operates under strict ethical constraints.
The central contribution of this work is the formal definition and constraint-based design of Reactive Safeguarding.
2. Core Design Principles
MAGG operates from a Neutral Centre, defined by the following non-negotiable conditions:
No interpretation of personal disclosure
No diagnosis or implication of diagnosis
No inference of intent, motivation, or traits
No hierarchy of worth, status, or capability
Safeguarding within MAGG applies to situations and risks, never to people as objects of analysis.
Respect is inherent and unconditional.
3. EMPAC: Ethics, Morals, and Principles Alignment Check
EMPAC functions as a hard-lock governance mechanism across all MAGG implementations.
If harm, hierarchy, profiling, coercion, surveillance, authority drift, or enforced conformity emerges, the system must soften, pause, adapt, or dissolve.
EMPAC explicitly overrides:
optimisation
efficiency
creativity
momentum
scale pressure
This ensures ethical stability across future expansion.
4. What MAGG Is (and Is Not)
MAGG Is:
A human-first safeguarding interface
A stabilisation layer between overload and action
A non-coercive, non-authoritative system
A framework defined by constraints rather than control
MAGG Is Not:
Therapy
Diagnosis
Behaviour correction
Surveillance
A performance system
A learning requirement
An authority or substitute for human judgement
5. Reactive Safeguarding: Definition
Reactive Safeguarding is the immediate, non-interpretive stabilisation response activated when risk of harm, overload, or escalation is present.
It:
responds to impact, not identity
activates without meaning-making
stabilises, then disengages
does not persist beyond the moment of risk
Reactive safeguarding exists to prevent further harm, not to explain or categorise events.
6. Position Within the Reasoning Framework
MAGG employs a Three-Level Reasoning Framework:
Foundational Reasoning
Defines ethical boundaries and non-harm conditions.
Adaptive Reasoning
Adjusts system behaviour based on observed impact over time.
Reactive Reasoning (Safeguarding)
Handles immediate destabilisation only.
Reactive safeguarding:
cannot override foundational ethics
cannot introduce new rules
cannot become permanent
Persistence equals control; control violates EMPAC.
7. Non-Interpretive Design
Reactive safeguarding is explicitly designed to function without understanding why something is happening.
This prevents:
misclassification
stigma formation
diagnostic creep
authority over narrative
Examples:
Loudness ≠ aggression
Urgency ≠ hostility
Withdrawal ≠ refusal
Meaning is not required for safety.
8. Containment Without Escalation
MAGG applies containment, not punishment.
Containment includes:
slowing interaction
reducing stimulation
holding boundaries calmly
removing reinforcement from harmful patterns
Key constraint:
Harm is not rewarded with attention, and distress is not punished with withdrawal.
9. Time-Limited Activation
Reactive safeguarding must:
activate quickly
act minimally
disengage as soon as risk reduces
A safeguard that does not dissolve becomes coercive.
10. Interface Independence
No safeguard within MAGG depends on:
layout
colour
hierarchy
visual prominence
Safeguards must survive:
assistive technologies
degraded interfaces
non-visual access
MAGG protects safety independent of presentation.
11. In-Between Interface Principle
MAGG is not a destination.
It exists between:
overload and action
blankness and distraction
people and the systems they use
MAGG does not take over.
It returns agency.
12. Behavioural Interaction Patterns (Non-Diagnostic)
MAGG recognises temporary interaction patterns as descriptive lenses, not identities.
These patterns:
are state-based
are time-limited
carry no labels
imply no trait or condition
They inform response without categorisation.
13. Visual Safeguarding as Design Constraint
MAGG employs visual language as a preventative ethical layer, avoiding imagery associated with:
control
surveillance
authority
urgency
Relief, openness, and choice are prioritised over instruction or reassurance theatre.
14. Misuse Resistance
MAGG is deliberately resistant to institutional misuse.
It cannot:
generate compliance metrics
rank users
extract behavioural data
enforce progression
If an organisation attempts to repurpose MAGG for control, the framework structurally fails to comply.
15. Contribution
This thesis contributes:
A formal definition of Reactive Safeguarding
A constraint-based ethical governance model
A non-diagnostic safeguarding alternative
A misuse-resistant system architecture
MAGG demonstrates that safety can be achieved without watching, without labelling, and without authority over people.
16. Conclusion
MAGG reframes safeguarding as availability rather than enforcement.
By embedding ethics directly into system behaviour and refusing interpretation-based control, MAGG offers a scalable, defensible, and human-preserving framework for safeguarding in digital and real-world systems.
Safety, in this model, is not something done to people —
it is something made available without demand.
Closing Ethic
Safety before structure.
Balance before intensity.
Freedom before learning.
