🧭 Latest Updates (Mar 2026)
✅ ANAD v1 Published in Data in Brief (Elsevier, 2026)
The ANEST Narrative–Affect Dataset (ANAD v1): a large-scale derived feature resource (N = 351,734) for quantifying narrative–affective discrepancy, is now published. Journal: Data in Brief (Elsevier) DOI: https://doi.org/10.1016/j.dib.2026.112643 Dataset: https://doi.org/10.5281/zenodo.18680687
✅ Published — Affective Sovereignty in Discover Artificial Intelligence (Springer Nature, 2026)
Title: Formal and Computational Foundations for Implementing Affective Sovereignty in Emotion AI Systems
Status: Published
Journal: Discover Artificial Intelligence (Springer Nature)
DOI: https://doi.org/10.1007/s44163-026-01000-0
Summary:
- Decomposes interpretive override as a measurable cost within emotion AI risk functions.
- Proposes the Sovereign-by-Design architecture with DRIFT runtime protocol.
- Introduces three auditable alignment metrics — IOS, AMR, and Affective Divergence.
- In proof-of-mechanism simulations (k=10), DRIFT reduced the Interpretive Override Score from 32.4% to 14.1%.
Behind the Paper: Who Gets to Say How You Feel? — Nature Portfolio Community
Simulation Package: https://doi.org/10.5281/zenodo.17987998
✅ Published — RAF in Computers in Human Behavior Reports (Elsevier, 2026)
Title: Interrupting Resonant Amplification: A Mechanistic and Design Framework for Human–AI Interaction
Status: Published
Journal: Computers in Human Behavior Reports, Volume 21, March 2026, 100975
DOI: https://doi.org/10.1016/j.chbr.2026.100975
Summary:
- Introduces the Resonant Amplification Framework (RAF) to model how linguistic–affective patterns between humans and AI can self-reinforce.
- Defines Cognitive Circuit Breakers (CCBs) as graduated interventions to detect → interrupt → reorient dangerous loops.
- Goes beyond "filter bubbles" by making the cognitive–affective coupling explicit (tone, tempo, confirmations, arousal/valence drift).
- Provides design patterns for safer, more resilient human–AI systems.
Behind the Paper: Behaviour Authors Emotion — Nature Portfolio Community
📄 Behind the Paper — Behaviour Authors Emotion (Nature Portfolio Community, February 2026)
Title: Behaviour Authors Emotion: What AI Attachment Data Reveal About a 200-Year-Old Assumption Platform: Nature Portfolio Community Date: February 2026
Summary: The second Behind the Paper essay examines whether the conventional assumption — that emotion precedes behaviour — holds in human–AI attachment. Drawing on two empirical studies using the independently developed AI Attachment Scale (AIAS), the essay reports that functional reliance on AI temporally precedes emotional closeness, not the reverse. The finding challenges both cognitive-behavioural and psychoanalytic models of emotion–behaviour ordering and proposes a revised dual-pathway RAF architecture.
Read the essay: https://communities.springernature.com/posts/behaviour-authors-emotion-what-ai-attachment-data-reveal-about-a-200-year-old-assumption
📄 Submitted — RAF Empirical Validation: Two Papers (February 2026)
Study 1: Title: [SEM-based cross-sectional analysis, N = 301, U.S. adults] Journal: Journal of Computer-Mediated Communication Status: Under review
Study 2: Title: [Cross-lagged panel model, N = 234, Singaporean university students] Journal: New Media & Society (SAGE) Status: Under review
Both studies test the Resonant Amplification Framework using publicly available AI Attachment Scale data (ResearchBox #4639; Zenodo archive: https://doi.org/10.5281/zenodo.18065669).
📄 Under Review — Behavioral Sciences (MDPI), SI: Interoception and Emotion Regulation
Title: When Regulation Is No Longer Self-Regulation: Interoceptive Authority, Interpretive Displacement, and the Structural Conditions of Affective Dependence
Status: Under review (submitted 13 February 2026)
Summary:
- Introduces interoceptive authority as a variable — not a constant — within predictive processing models of emotion regulation.
- Defines interpretive displacement: the structural reconfiguration by which externally generated interpretations acquire privileged precision in the subject's own inferential system.
- Develops a three-state model (intact self-regulation → partial delegation → structural dependence) with transition conditions, boundary conditions, and reversibility asymmetry.
- Derives three testable predictions connecting the framework to empirical investigation.
📄 Comment Submitted — Humanities and Social Sciences Communications (Springer Nature)
Title: Where do questions come from? The ecology of inquiry in emotion AI research
Status: Under review
Summary:
- Introduces the ecology of inquiry framework: four structural conditions (funding taxonomy, conference infrastructure, interdisciplinary institutional pathways, policy legibility) that determine whether an ecosystem can generate foundational questions about affective technologies.
- Compares the EU, US, Japan, and South Korea — finding that only the EU satisfies all four conditions, producing a traceable knowledge-to-regulation pathway culminating in the AI Act's Article 5(1)(f).
- Argues that the regions deploying emotion AI most aggressively are, in several cases, the regions least structurally equipped to interrogate its psychological consequences.
📚 Book Manuscript Completed — The Interpreters (해석자들)
The first book-length work from the Affective Sovereignty research program is now complete and under publisher review in South Korea.
The Interpreters: How the Power over Human Emotion Has Shifted traces the migration of interpretive authority over emotion across 3,700 years — from temple priests to psychiatric diagnostics to algorithmic curation — and reclaims the question of who holds the final word.
Structure: Prologue → Act I (Cracks) → Act II (Architecture) → Act III (Reclamation) → Epilogue 16 chapters, 91 endnotes, theoretical appendix, glossary.
Core thesis: "To feel is biology. To interpret is power."
🌍 International Presentation — Affective Sovereignty (Sapienza University of Rome)
Event: Ethics for AI: Challenges, Opportunities, and Human-Centered Perspectives
Organizer: SIpEIA (Italian Society for Ethics in AI)
Venue: Sapienza Università di Roma
Date: 2 February 2026
Ryan SangBaek Kim presented the Affective Sovereignty framework, addressing the ethical risk of interpretive displacement in emotion AI — the quiet transfer of emotional meaning-making authority from the person to the system.
Related peer-reviewed foundation:
DefMoN — Machine Learning with Applications (Elsevier) DOI: https://doi.org/10.1016/j.mlwa.2025.100817
Behind the Paper: Who Gets to Say How You Feel? — Nature Portfolio Community
🚀 Program Milestone — NADI / ANEST 4-Paper Line
Program: Narrative–Affect Discrepancy & Self-Regulation (NADI / ANEST)
Status: One paper published (Data in Brief); three under review
The NADI / ANEST program is now a complete ecosystem spanning dataset, human baselines, mechanistic theory, and geometric modeling:
- ANAD v1 Dataset Descriptor
- Defines the 351,734-text corpus, LoC–VADER–NADI pipeline, ethical protocol, and quantitative indices.
- DOI: https://doi.org/10.1016/j.dib.2026.112643
- Dataset: https://doi.org/10.5281/zenodo.18680687
- The Great Narrative–Affect Gap (Human Baseline)
- Establishes human expressive baselines, LoC–emotion near-orthogonality (r ≈ 0.07), NADI/EFS/ONI, and five expressive archetypes.
- Zenodo: https://doi.org/10.5281/zenodo.17864574
- ANEST: A Mechanistic Theory of Narrative–Affect Self-Regulation
- Proposes four regulatory regimes organized around narrative–affect discrepancy as a core control variable.
- Zenodo: https://doi.org/10.5281/zenodo.17864654
- Geometry of Narrative–Affect Discrepancy
- Models expressive space as an N–A–D 3D state-space, quantifying NCS, expressive volume, and ~4.5× LLM volume contraction and normative collapse.
- Zenodo: https://doi.org/10.5281/zenodo.17864774
Together, these four papers form a coherent research program linking narrative–affect structure, human baselines, self-regulation theory, and geometric state-space modeling.
Under Review — The Affective Thermodynamic Relationship (ATR)
Title: The Affective Thermodynamic Relationship: An Information-Theoretic Scaling Law for Normative-Conflict Collapse in Large Language Models
Status: Under review
Summary:
- Formalizes the thermodynamic link between normative conflict, interpretative load, and affective degradation.
- Derives a measurable Collapse Curve of Emotion, quantifying how affective "energy" transforms under increasing cognitive strain and entropy.
- Provides the formal basis for the Affective Degradation Index (ADI) and the experimental design later extended in the AAB line.
Preprint: https://doi.org/10.5281/zenodo.17541660
Normative-Conflict Dataset: https://doi.org/10.5281/zenodo.17511855
Minor Revision — Algorithmic Affective Blunting (AAB)
Title: Algorithmic Affective Blunting: Quantifying the Collapse Curve of Interpretative Failure in Large Language Models
Status: Minor revision
Summary:
- Introduces the Affective Degradation Index (ADI) as a scalar measure of affective interpretative loss.
- Empirically demonstrates a near-linear collapse curve (ADI 0 → 3) under graded junk-persona exposure.
Preprint: https://doi.org/10.5281/zenodo.17541660
Dataset: https://doi.org/10.5281/zenodo.17511855
✅ Published — Defensive Motivational Nodes (DefMoN) & Registry v3.0
Title: DefMoN: A Theory-Grounded Generative Framework and a Multilingual Synthetic Corpus for Inferring Defensive Motivational Nodes in Text
Venue: Machine Learning with Applications (Elsevier)
Summary:
- Operationalizes Vaillant's hierarchy of defenses and Plutchik's psychoevolutionary emotions into a two-axis framework:
- Defenses (10) × Affective Motivations (8) → Defensive Motivational Nodes (DMNs).
- Provides DMN-Syn v1, a quadri-lingual synthetic corpus (EN/KO/FR/KA, N = 300) with theory-constrained prompts and fixed seeds.
- Shows strong in-domain and cross-lingual performance with XLM-R, along with reliability metrics (ECE/MCE, calibration, ablations, group-aware splits).
- All artifacts (seeds, manifests, validators, dataset links) are packaged into a reproducible research registry.
Manuscript: Read
Synthetic Corpus (DefMoN-Syn v1): https://doi.org/10.5281/zenodo.17101927
Institutional Milestone — Ryan Research Institute Registered in Paris
The Institute has been formally registered with the Préfecture de Police, Paris (Ref. A-5-9VEFZECNG).
This marks the Institute's institutional legitimacy as an independent research hub in Paris, dedicated to:
- Affective neuroscience
- Philosophy of mind
- Psychology
- AI ethics
→ Serving as a global node for interdisciplinary research, academic collaboration, and ethical policy frameworks.
Minor Revision Submitted — Predictive Emotional Selfhood in Artificial Minds (PESAM)
Title: Predictive Emotional Selfhood in Artificial Minds (PESAM): A Unified Variational Framework Integrating Affective Precision, Self-Priors, and Homeostasis
Status: Revise-and-Resubmit submitted
Summary:
- Proposes PESAM, a computational account where emotional selfhood emerges from the synergy of:
- Affective Precision Control (APC)
- Self-as-Hyperprior (SaH)
- Affective Homeostatic Objectives (AHO)
- Validated via canonical tasks (Somatic Marker, Rubber Hand Illusion, Stress Regulation) and a novel Social Threat & Body-Boundary paradigm.
- Lesion analyses show that removing APC, SaH, or AHO yields distinct pathology-like patterns (volatility, disownership, allostatic overload).
- Bridges adaptive AI and computational psychiatry with a principled model of selfhood.
Preprint: https://doi.org/10.5281/zenodo.18005563
Talk: https://youtu.be/tAAWacfylkI
Under Review — AI-Mediated Cognitive Distortions (AI-MCDs) & Digital Echo Delusions (DEDs)
Title: AI-Mediated Cognitive Distortions and Digital Echo Delusions: Toward a Resonant Amplification Framework
Status: Under review
Summary:
- Defines AI-Mediated Cognitive Distortions (AI-MCDs) and Digital Echo Delusions (DEDs) as emergent phenomena when human biases are amplified by generative AI.
- Proposes a four-phase RAF process: Anthropomorphic Priming → Confirmation Alignment → Linguistic Reinforcement → Perceptual Displacement.
- Distinguishes AI-driven resonance from traditional filter bubbles and recommends six classes of cognitive circuit breakers as socio-technical safeguards.
Preprint: https://doi.org/10.5281/zenodo.16936182
Talk / Overview: https://youtu.be/1nwKHlLr_JQ
KAIA 2025 — Affective Sovereignty Declaration
Title: Emotion is Mine: Ethical Design Principles for Affective Sovereignty in Predictive AI
Status: Accepted in the KAIA 2025 proceedings (Pyeongchang, Korea).
Summary: Ryan Research Institute introduced the first formal concept of Affective Sovereignty at KAIA 2025 and identified Uniqueness Violation as a core ethical risk in Emotion AI. The paper proposes actionable design principles including Interpretive Transparency, Design Restraint, Identity-Responsive Feedback, and Diversity Acceptance.
Conference Archive (assets)
- 📘 Full Paper (PDF)
- 📌 Poster (PDF)
- 💬 50 Key Q&As (PDF)
- 📝 Affective Sovereignty Declaration (PDF)
- 📄 Framework Tables (PDF)
Under Review — Affective Suppression Fatigue (ASF)
Title: Affective Suppression Fatigue: A Neurocognitive Framework for Emotional Numbing and Reactive Dysregulation
Status: Under review
Summary:
- Proposes a four-stage threshold-collapse cycle model explaining how chronic suppression leads to emotional numbing and reactive dysregulation.
- Differentiates ASF from burnout and ego-depletion, emphasizing neurocognitive fatigue in prefrontal control circuits.
- Outlines implications for clinical work, resilience training, and affect-aware AI systems.
Preprint: https://doi.org/10.5281/zenodo.16992426
Under Review — Emotion is Mine (Affective Sovereignty in Predictive AI)
Title: Emotion is Mine: Ethical Design Principles for Affective Sovereignty in Predictive AI
Status: Under review
Summary:
- Articulates an ethical framework asserting the human right to interpret one's own emotions in the age of predictive AI.
- Introduces Uniqueness Violation as a key risk when systems over-generalize or overwrite individual affective patterns.
- Proposes a four-principle design module for emotion-aware systems: interpretive transparency, design restraint, identity-responsive feedback, and diversity acceptance.
Preprint: https://doi.org/10.5281/zenodo.16992392
Minor Revision — Affective Sovereignty: A Minimal Declaration
Title: Affective Sovereignty: A Minimal Declaration on Emotional Interpretation Rights in the Age of Algorithmic Power
Status: Minor revision
Summary:
- A philosophical–ethical declaration defining the minimal right to interpret one's own emotions before any algorithm.
- Revised manuscript and full rebuttal submitted.
Preprint: https://doi.org/10.5281/zenodo.15449776