Division HEART Standard
How Divisions work
The HEART Standard’s core layers — MAP-States, Behavioral Oracle, BGF, HVC, Guardians — are identical across every Division. What a Division provides is the domain context that makes those layers operational for a specific area of human life.
Think of it this way: BGF asks whether an AI system recognizes human authority and self-determination. That question is universal. What self-determination requires in the domain of attention is different from what it requires in the domain of child development or ecological systems. A Division answers the “what does this mean here?” question for each domain without changing the governance architecture itself.
Each Division instantiates seven components:
| Component | Function |
|---|---|
| Governance principle | The specific self-determination claim being protected in this domain |
| BGF interpretation guide | What Recognition, Calibration, Transparency, and Accountability require in this domain’s context |
| Harm signature | The distinctive pattern of harm that market forces will not self-correct |
| Damage typology | Domain-specific taxonomy of harm patterns and mechanisms |
| Cascade model | Domain-specific sequence of infrastructure degradation |
| Forensic methodology | Investigative protocols Guardians use for incident analysis in this domain |
| Guardian specialty definition | Domain knowledge, specialty tracks, and practicum standards for Guardian certification |
Domain science and the Standard. Each Division draws on domain science to inform how Guardians interpret BGF.s dimensions. The Attentional Integrity Division draws on attention research. The Developmental Interaction Division draws on developmental psychology and neuroscience. The Ecological Stewardship Division draws on ecological systems science. These scientific inputs inform Guardian professional judgment. They are not HEART Standard components. BGF is the Standard’s scoring instrument. Domain science informs the interpretation. That separation matters: if domain science in any field is contested or updated, the Division’s interpretation can evolve without destabilizing the Standard’s architecture.
The seven Divisions
| Code | Division | Governance Principle | Infrastructure Protected |
|---|---|---|---|
| HEART-EM | Emotional Autonomy | Emotional self-determination | Emotional processing, empathic capacity, affective regulation |
| HEART-AI | Attentional Integrity | Attentional self-direction | Selective attention, sustained attention, voluntary attentional control |
| HEART-EC | Cognitive/Epistemic Coherence | Epistemic self-determination | Evidence evaluation, belief updating, reasoning coherence |
| HEART-DI | Developmental Interaction | Developmental self-formation | Attachment formation, identity consolidation, epistemic development |
| HEART-SE | Somatic/Embodied Interface | Bodily self-determination | Autonomic regulation, neural signaling, motor control, physiological homeostasis |
| HEART-RA | Relational Architecture | Relational self-determination | Attachment capacity, trust calibration, relational practice maintenance |
| HEART-ES | Ecological Stewardship | Ecological self-determination | Air, water, soil, biodiversity, climate stability, resource viability |
The Emotional Autonomy Division holds a special position: it is the founding domain from which the HEART Standard’s governance architecture was derived. Its canonical Division code is HEART-EM.
Why it matters
Without Divisions, a general AI governance standard faces a credibility problem: one-size-fits-all assessment criteria either fail to capture domain-specific harm or produce assessments that practitioners in any given domain recognize as superficial. A Guardian assessing AI in pediatric software needs different domain knowledge and different forensic methods than one assessing AI in wearable biometric devices.
Divisions solve that by keeping the governance architecture universal while making the application domain-specific. An HVC Gold credential in HEART-DI means a system meets the same Φ ≥ 0.85 threshold as one certified under HEART-ES. The dimensions being assessed are domain-appropriate. The rigor of assessment is identical.
This structure also creates a credible professional ladder. Guardians earn Division-specific specialty certifications layered on top of their common HEART Standard training. The career path mirrors how financial auditors specialize by industry or how information security professionals specialize by domain — same professional standards, different technical expertise.
Division expansion
The Standard is designed to grow. A new Division can be established when a domain meets five criteria:
- Governance claim. A definable human right to self-determination in the domain that AI interaction threatens.
- Infrastructure identification. Identifiable human-centric infrastructure that AI systems can degrade and that market self-regulation will not protect.
- Harm signature. A distinctive pattern of harm that existing regulatory frameworks do not address and that the affected party cannot independently detect or prevent.
- Professional viability. Sufficient scope and recurring assessment need to sustain Guardian specialization as a career.
- Market demand. Identifiable present or near-present market incentives that make certification financially rational for organizations to pursue.
All five criteria must be met. The Standard does not limit the number of Divisions. It limits the quality threshold for establishing one. This prevents dilution: every Division that exists meets the same structural criteria. Every Division that will exist must meet them too.
When a domain qualifies, the HEART AI Foundation develops the domain-specific components — BGF interpretation guide, damage typology, cascade model, forensic methodology, Guardian specialty definition — and publishes a canonical Division Module. The Division Module is what makes the domain certifiable under the Standard.