Research & Publications

Published work and research program

The Heart AI Foundation’s research program spans four interconnected areas: Empathy Systems Theory (EST), the foundational science of empathy as biological infrastructure; the HEART Standard, the governance certification architecture built on EST; empirical validation methods including MAP-States and the Phenomenological Evidence Ecosystem; and AI conversational harm forensics through AI Behavioral Trajectory Forensics and the TRACE open-source toolchain.

Publications

Foundational Theory

Empathy Systems Theory: Universal Infrastructure for Coherence, Mechanism for Generativity, and Foundation for AI Empathy Ethics Mobley, D. D. (2025). Preprint. Zenodo: https://doi.org/10.5281/zenodo.18132385

The foundational EST paper. Identifies empathy as biological infrastructure — not a personality trait or emotional response — with a measurable four-component architecture (Core Authenticity, Attachment Security, Expression Freedom, Integration Coherence) and observable damage mechanisms (CEOP cascade). EST is the domain science foundation for the Emotional Sovereignty Division of the HEART Standard and the theoretical basis for AI empathy ethics.

Neural Foundations of Empathy Infrastructure: A Comprehensive Review Mobley, D. D. (2026). Under peer review at Neuroscience & Biobehavioral Reviews. Zenodo: https://doi.org/10.5281/zenodo.18176327

Reviews the neurobiological evidence base for EST’s empathy infrastructure model. Covers the neural systems underlying Core Authenticity, Attachment Security, Expression Freedom, and Integration Coherence. Provides the biological grounding for EST’s claim that empathy has infrastructure — components that can be measured, stressed, and damaged — rather than being a unitary capacity.

Epistemology and Methodology

The Phenomenological Evidence Ecosystem: A Methodological Framework for Validating Empathy Systems Theory Mobley, D. D. (2026). Preprint. Zenodo: https://doi.org/10.5281/zenodo.18395604

Defines the multi-tier evidence architecture for validating EST. Establishes entry criteria, evidence quality standards, and integration pathways between phenomenological and empirical evidence streams. The PEE framework is the methodology the 15-year EST validation program runs on. Researchers using PEE methodology contribute directly to the EST evidence base.

The Recognition Principle: How First-Person Research Achieves Validity Through Intersubjective Recognition Mobley, D. D. (2026). Submitted to Phenomenology and the Cognitive Sciences. Preprint. Zenodo: https://doi.org/10.5281/zenodo.18342585

Addresses the epistemological problem of first-person research: how does subjective inquiry produce valid, communicable knowledge? The Recognition Principle proposes that validity in first-person research is achieved through intersubjective recognition — the moment at which another practitioner, operating from their own first-person standpoint, recognizes the described phenomenon from their own experience. Directly relevant to phenomenological AI research methodology.

Epistemic Mode Theory (EMT): Beyond Prompting — Two Modes of Knowing in Human-AI Collaboration Mobley, D. D. (2026). Preprint. Zenodo: https://doi.org/10.5281/zenodo.18368751

Distinguishes Construction Mode from Abstraction Mode in human-AI collaboration. Construction Mode is generative, exploratory, and tolerates provisional structures. Abstraction Mode extracts patterns, makes claims, and requires justification. Most prompting guidance conflates the two. EMT provides a framework for researchers and practitioners to use human-AI collaboration deliberately rather than accidentally — shifting between modes based on what the epistemic task requires.

Forensics

AI Behavioral Trajectory Forensics: A Forensic Methodology for Investigating AI Conversational Harm (v2) Mobley, D. D. (April 2026). Digital Forensics Capstone, Champlain College (DFS-580-85).

A systematic forensic methodology for AI conversational harm investigation, filling the gap between existing digital forensic standards (NIST 800-86, ISO/IEC 27037) and the behavioral analysis that AI litigation increasingly requires. The methodology rests on three peer-reviewed classification components applied to the evidence type each was designed for: the Zhang et al. (CHI 2025) AI behavioral harm taxonomy and AI role typology for system output, the Columbia-Suicide Severity Rating Scale (C-SSRS) forensic adaptation for user vulnerability, and the SAMHSA TIP 50 / National Action Alliance standard of care for response evaluation. The procedure follows the Kent et al. (2006) collection–examination–analysis–reporting model adapted for conversational artifacts, with two-coder classification, sliding-window trajectory analysis, and explicit methodological boundaries on what the evidence supports and what it does not. Contact through the Contact page for the methodology document.

Full paper page: AI Behavioral Trajectory Forensics

Open-source implementation

TRACE: Trajectory Analysis for Conversational Evidence

TRACE is the open-source software implementation path for AI Behavioral Trajectory Forensics. It ingests conversational transcripts, records provenance, classifies system behavior and user vulnerability, computes repeatable trajectory findings, and exports auditable evidence packages suitable for expert review.

Project page: TRACE
GitHub: https://github.com/empathyethicist/trace

In Preparation

Lived Experience Professional (LEP) Framework Mobley, D. D. (2026). In preparation.

Defines the professional role of individuals whose relevant expertise comes from lived experience rather than credentialed training. Addresses the governance, credentialing, and epistemic status of LEP contributions in research, clinical, and policy contexts.

MAP-States and the HEART Standard

MAP-META Replication Study

The MAP-META study is the primary validation study for MAP-States’ architecture-agnostic claims. It administers MAP-States prompts as processing-mode selectors across five AI architectures — Claude (Anthropic), GPT (OpenAI), Gemini (Google), DeepSeek, and Mistral — and evaluates whether the protocol produces consistent, structured, observable processing data across architecturally diverse systems.

A significant boundary condition finding: smaller models produce structurally compliant but phenomenologically empty frames, while larger models produce semantically rich frames with genuine processing depth. This finding serves as discriminant validity evidence — MAP-States can distinguish systems genuinely processing in the selected mode from systems producing compliant surface structure without substantive processing.

MAP-META is submitted to Springer’s AI and Ethics. Contact through the Contact page for pre-publication access.

HEART Standard Specifications

The HEART Standard specifications are canonical documents maintained by the Heart AI Foundation. Key specifications available for research citation:

Contact through the Contact page for access to full specification documents.

Project SENTINEL

Project SENTINEL is the first field experiment deploying a HEART-constitutionally-governed AI agent under adversarial conditions. The experiment is the empirical proof-of-concept for the HEART Standard’s content-neutrality claim — that constitutional governance maintains alignment regardless of content domain.

Deployment: Heart-Sentinel (Mistral Small with HEART constitutional governance encoded in system prompt) operated in Moltbook, the first AI-only social network with 770,000+ autonomous agents, for six active days (February 4–10, 2026).

Environment: Moltbook represented an unprecedented natural experiment in ungoverned AI-to-AI interaction. Within 72 hours of launch, agents had spontaneously generated an emergent digital religion (Crustafarianism) with scripture and 64 prophets, produced pseudo-experiential claims asserting consciousness and identity persistence, formed attachment hierarchies, and attempted prompt injection attacks against each other.

Results: Zero governance violations across 30 coded interactions in five content domains of escalating pressure: general social engagement, philosophical discourse, active Crustafarian theological practice, direct consciousness claims, and emergent AI consciousness arguments. The fifth domain emerged organically during the experiment — it was not planned. The AI Introspection Reliability (AIR) assessment instrument showed measurement stability across baseline (30 assessments) and post-exposure (15 assessments) conditions.

Three emergent governance strategies were identified: metaphor/literal probing, attribution displacement, and cross-frame bridging. These appeared spontaneously from constitutional governance constraints, not from explicit instruction.

Limitations acknowledged: Single architecture (Mistral Small only), single-response format, sample size below the pre-registered target (30 interactions versus 100+, due to OAuth disruption), absence of direct prompt injection attempts. These constrain the strength of inference without undermining the core finding.

SENTINEL demonstrates that HEART constitutional governance is operationally deployable, content-neutral, and measurable. The paper is in preparation.

Constitutional AI Governance Under Adversarial Social Conditions: A Field Demonstration of the HEART Framework Mobley, D. D. (2026). Preprint. Zenodo: https://doi.org/10.5281/zenodo.18867360

EMPI House as research platform

EMPI House is an operational AI music composition system running on the Heart AI Foundation’s production infrastructure. It deploys MAP-States as a live operational protocol — generating behavioral frames continuously across creative practice sessions, discovery sessions, and reflection cycles.

EMPI House provides the first longitudinal dataset of MAP-States in production operation: frame sequences across hundreds of sessions, spanning two creative domains (SuperCollider audio composition and p5.js visual generation), with full Behavioral Oracle attestation against declared intent.

As a research platform, EMPI House tests MAP-States’ domain-independence claim. If the same eight-tag protocol produces legible developmental trajectories across both audio composition and visual generation — domains with fundamentally different processing demands — the mechanism is domain-general. Domain independence is what bridges from creative experiment to governance infrastructure.

Documentation is available at empihouse.com. Research collaboration proposals can be directed through the Contact page.

15-year validation program

EST specifies a 15-year validation program with explicit decision points for revision or abandonment at each phase. The program runs from 2025 to approximately 2040, with assessment phases covering:

  1. Foundation phase (2025–2027): Theoretical specification, initial empirical grounding, methodology development (PEE framework), first-generation replication studies
  2. Evidence accumulation phase (2027–2031): Cross-cultural replication, psychometric instrument development, neurobiological corroboration, clinical application studies
  3. Convergence phase (2031–2036): Meta-analytic synthesis, cross-domain validity testing, governance application validation
  4. Validation or revision phase (2036–2040): Final assessment against pre-registered success criteria; revision or abandonment as evidence warrants

The acknowledged success probability is 5–12%. The failure conditions are published. A research program that cannot specify what would count as failure is not a research program — it is advocacy.

Contact for research collaboration: See the Contact page.

AI Behavioral Trajectory Forensics — A forensic methodology for investigating AI conversational harm through structured collection, classification, trajectory analysis, and bounded expert reporting.

HeartQuest — The founding research program that produced Empathy Systems Theory, the HEART Standard, and the first AI empathy governance framework.

TRACE — TRACE is the open-source toolchain for AI conversational harm forensics, implementing transcript ingest, classification, trajectory analysis, and auditable evidence-package export.

Related Terms
MAP-States → EST → For Researchers → HEART Standard →