Newsletter

Research essays, public scholarship, and media writing from Ryan Research Institute.

This page gathers work on Affective Sovereignty, interpretive authority, Mental AI, human–AI interaction, AI evaluation and governance, authorship and provenance, and philosophy of mind. It separates research-led essays from broader public writing while preserving links to the underlying publications where relevant.

Long-form essays are published primarily on Substack. Research narratives around peer-reviewed papers also appear through Springer Nature Research Communities and selected public media.

Subscribe here: https://substack.com/@profryankim

Latest Research Writing

The Curve Was Stable. The Judgment Was Not.

Springer Nature Research Communities · Behind the Paper · 22 August 2026

A reflection on the Communications AI & Computing paper on the Affective Thermodynamic Relationship, focusing on why a stable empirical collapse curve can coexist with substantial evaluator disagreement.

→ Springer Nature Research Communities profile

Can an AI Evaluate a Mind It Has Already Helped Change?

Springer Nature Research Communities · Behind the Paper · 5 August 2026

On post-deployment criterion endogeneity and why later agreement may cease to be independent evidence once an AI inference has entered a person’s self-understanding, behaviour, or institutional record.

→ Read

Can Claude’s Watermark Prove AI Authorship?

Substack · August 2026

On the distinction between technical provenance, intellectual contribution, and authorship, with translation as a boundary case and institutional use of detection as the governance problem.

→ English · → Korean edition

The Human Mind Is Becoming AI’s Operating Environment

Substack · 3 August 2026

A public explanation of Mental AI and post-deployment criterion endogeneity as a causal problem in AI evaluation.

→ English · → Korean edition

Springer Nature Research Communities · Behind the Paper

• The Curve Was Stable. The Judgment Was Not. — 22 Aug 2026
• Can an AI Evaluate a Mind It Has Already Helped Change? — 5 Aug 2026
• Affective Sovereignty and the Epistemic Gap: From System Design to Measurement Theory — 12 Jul 2026
• How Emotional AI Fails Without Sounding Broken — 6 Jul 2026
• How Emotional Discrepancy Stopped Looking Like Error — 13 May 2026
• When Good Design Becomes the Problem: Reflective Summarization and the Measurement Gap in Affective AI — 13 Apr 2026
• Interpretive Authority: The Missing Layer in AI Governance — 6 Apr 2026
• Can Self-Awareness Exist Without a Self? What Indian Philosophy Asks of Consciousness Science and AI — 11 Mar 2026
• Affective Preconditions for AI Safety: What the Anthropic-Pentagon Dispute Reveals — 7 Mar 2026
• Behaviour authors emotion: What AI attachment data reveal about a 200-year-old assumption — 28 Feb 2026
• Who Gets to Say How You Feel? — 18 Feb 2026

→ Full Springer Nature Research Communities profile

Selected 2026 Substack Research Essays

• Can Claude’s Watermark Prove AI Authorship? · Korean edition
• The Human Mind Is Becoming AI’s Operating Environment · Korean edition
• The Risk Is Not the Reply. It Is the Relationship.
• The Machine Is Certain. That Does Not Mean It Knows What You Feel.
• Your AI Can Read You. It Should Never Outrank You.
• Your AI Sounds Like It Cares. That’s the Problem.
• Korean essay: How a Quiet Mind Avoids Collapse
• The Missing Coordinate in the RLHF Diversity Debate
• Stop Naming Your Emotions
• Who Gets to Say How You Feel?
• Does Emotion Cause Behaviour — or Does Behaviour Create Emotion?
• When Accuracy Is Not the Problem

→ Full Substack archive

Featured Essays

Interpretive Authority: The Missing Layer in AI Governance

Springer Nature Research Communities, April 2026

Governance of emotion AI requires more than privacy, bias, or transparency. It must also confront the question of who has the authority to define emotional meaning.

Who Gets to Say How You Feel?

Springer Nature Research Communities

A Behind the Paper essay on Affective Sovereignty and the problem of interpretive override in emotion AI.

Behaviour Authors Emotion — Historical Behind the Paper

Springer Nature Research Communities

A 2026 Behind the Paper essay on human–AI attachment and the early empirical interpretation behind the Resonant Amplification Framework. Its directional interpretation has since been superseded by corrected public analysis; current RRI evidence does not establish that social substitution precedes emotional closeness or the reverse.

Emotion AI and Interpretive Authority

Where Emotion AI Actually Breaks

Why emotion AI fails not only at the level of models or data, but at the level of hidden assumptions about interpretation.

The Ethics of Emotional Inference

On the limits of emotional inference in AI, and why affective sovereignty matters when systems begin to name, predict, or respond to feeling.

Affective Sovereignty: A Minimal Declaration

A short statement on emotional interpretation rights in the age of algorithmic power.

GPT-6 Remembers You. But Who Owns Your Emotions?

On memory, intimacy, and the question of emotional ownership in AI systems.

Research Notes and Concept Essays

Can Claude’s Watermark Prove AI Authorship?

On the difference between technical provenance and intellectual contribution, why translation is a revealing boundary case, and how watermark evidence may change once institutions make detection consequential.

Korean edition

DefMoN: Reading the Hidden Defenses in Language

How language encodes defensive patterns, and why that matters for empathy, safety, and affective AI.

The Affective Thermodynamic Relationship

A research note on normative-conflict collapse, information-theoretic strain, and the possibility of lawful scaling in large language model failure.

Related article: Communications AI & Computing, published 2026 · Version of Record live

DOI: https://doi.org/10.1038/s44488-026-00006-y

Algorithmic Affective Blunting

How large language models lose affective interpretive coherence under semantic stress, and why emotional AI evaluation should move from apparent empathy to interpretative robustness.

Related article: Discover Artificial Intelligence, published (2026)

DOI: https://doi.org/10.1007/s44163-026-01573-w

Mental AI: When AI Inferences Change Their Own Evidence

On post-deployment criterion endogeneity: how AI inferences can alter the human states, evidence, and criteria later used to evaluate those same inferences. Public preprint DOI 10.20944/preprints202607.2118.v1.

PESAM: Predictive Emotional Selfhood in Artificial Minds

On selfhood, affective priors, and emotional regulation in artificial systems.

Essays on Selfhood, Culture, and Society

When AI Becomes a Partner in Your Mind

An introduction to the Resonant Amplification Framework and the co-evolution of human thought and machine response.

The Hidden Cost of Suppression

An essay on Affective Suppression Fatigue and the structure of emotional numbing.

Doubt, Faith, and the Moment of Contact

A philosophical reading of Caravaggio and the structure of belief, touch, and uncertainty.

Under the Light That Conceals Truth

A cultural and civilizational reflection on contemporary Korea.

Media and Public Writing

AI Watermarks Can Be Accurate While Authorship Judgments Are Wrong

AI Times Korea, 19 August 2026

A contributed essay distinguishing technical provenance from intellectual contribution and authorship, and arguing that watermark or detector signals should not acquire adjudicative force by themselves.

→ Read at AI Times

The Missing Right in High-Impact AI: The Ability to Correct an Interpretation

AI Times Korea, 11 August 2026

A policy essay translating interpretive authority into operational requirements for high-impact AI: notice, uncertainty and alternatives, abstention, subject correction, correction propagation, and meaningful human review.

How Emotion AI Misses the Structure of Human Feeling

MIT Technology Review Korea, May 2026

A public essay on why emotion AI may fail not by misreading emotions, but by interpreting them too quickly, before the person has had time to form their own emotional meaning.

A New Study Mapped 350,000 Relationship Stories and Found a Communication Style AI Struggles to Copy

PsyPost, May 2026

Media coverage of the PLOS ONE study on narrative–affect discrepancy, strategic understatement, expressive collapse, and the limits of aligned AI in reproducing quieter forms of human emotional expression.

Why Does AI Name Your Feelings Before You Do?

MIT Technology Review Korea, April 2026

A public essay on reflective summarization, affect labeling, and the risk that helpful AI responses may begin to occupy the first interpretive position in emotional life.

Can Humans Retain Interpretive Authority over Their Own Emotions?

MIT Technology Review Korea, March 2026

A public essay on affective sovereignty, alignment faking, and the need to protect the human capacity to interpret one’s own emotional life.

The Real Risk of Emotion AI Is Not Accuracy, but the Transfer of Interpretive Power

AI Times, February 2026

A public essay arguing that the core risk of emotion AI is not merely misclassification, but the structural transfer of interpretive authority from people to systems.

For the full archive of essays and ongoing writing, visit the Full Substack Archive.