/* Books 페이지 카드 클릭 비활성화 */ a.notion-collection-card { pointer-events: none !important; cursor: default !important; }

Presentations

Affective Sovereignty at Sapienza Università di Roma

International Conference Presentation

image
image

Sapienza Università di Roma, Faculty of Letters and Philosophy — SIpEIA 2026

Event Overview

Title of Talk

Affective Sovereignty: Reclaiming the Right to Feel for Oneself

Conference

Ethics for AI: Challenges, Opportunities, and Human-Centered Perspectives

Session

Accountability and Care

Organizer

SIpEIA — Italian Society for Ethics in Artificial Intelligence

Host Institution

Sapienza Università di Roma

Faculty of Letters and Philosophy

Location

Rome, Italy

Date

2 February 2026

Presenter

Ryan SangBaek Kim, Ph.D.

Founder & Director, Ryan Research Institute (Paris)

Context of the Presentation

The presentation was delivered as part of the international SIpEIA 2026 conference, an interdisciplinary forum bringing together scholars, policymakers, technologists, and ethicists to address emerging challenges in artificial intelligence from a human-centered perspective.

The “Accountability and Care” session focused on ethical responsibility, interpretive authority, and relational dimensions of AI systems deployed in sensitive domains such as hiring, mental health, education, and companionship technologies.

Within this context, the talk introduced Affective Sovereignty as a conceptual and normative framework addressing a specific but underexamined ethical risk in emotion AI systems.

Central Thesis

The core argument of the presentation was that the most serious ethical risk posed by emotion AI is not misclassification, but the structural displacement of interpretive authority.

When algorithmic emotional assessments become records, defaults, or decision anchors, the system quietly assumes the role of final interpreter—reducing the human subject from an interpreting agent to a data point.

This phenomenon was termed interpretive displacement.

The presentation argued that:

  • Privacy frameworks protect emotional data
  • Fairness frameworks regulate outcomes
  • Affective Sovereignty protects meaning

Key Concept: Interpretive Displacement

Interpretive displacement occurs when:

  • Emotional labels generated by systems are treated as authoritative
  • Human self-interpretation becomes secondary or corrective rather than primary
  • Individuals lose the practical right to contest how their emotions are defined

This shift does not require system error.

It can occur even when predictions are statistically accurate.

The ethical problem arises when accuracy replaces agency.

Core Claims Presented

  1. Emotion is not merely observable data

    2. Emotional experience is interpretive, contextual, and historically situated.

  2. Interpretation is a form of authority

    3. Whoever has the last word on emotional meaning holds power over identity, memory, and action.

  3. Emotion AI systems currently lack interpretive safeguards

    4. Most systems optimize prediction without protecting the user’s right to reinterpret or refuse interpretation.

  4. Affective Sovereignty defines a minimal ethical boundary

    5. Before any algorithmic interpretation, the person retains the right to define, revise, or reject emotional meaning.

Design Principles Proposed

The presentation outlined three auditable design principles for emotion-aware AI systems:

1. Interpretive Restraint

Systems must limit claims about emotional meaning to probabilistic, non-final representations.

2. Interpretive Provenance

All emotional inferences must disclose how, why, and under which assumptions interpretations were produced.

3. User Interpretive Authority

Users must retain the right to contest, override, or reinterpret emotional outputs without penalty.

These principles were framed as design constraints, not optional UX features.

Relation to Ongoing Research

The presentation was grounded in a broader research ecosystem developed at the Ryan Research Institute:

  • Defensive Motivational Nodes (DefMoN)

    • A theory-grounded framework linking affective motivation and defense mechanisms in language.

    Published in Machine Learning with Applications (Elsevier, 2026).

    DOI: https://doi.org/10.1016/j.mlwa.2025.100817

  • NADI / ANEST Program

    • A four-paper research line integrating dataset construction, human baselines, mechanistic theory, and geometric modeling of narrative–affect discrepancy.

  • Algorithmic Affective Blunting (AAB)

    • Empirical measurement of affective degradation and collapse curves in large language models.

These works provide both theoretical and empirical grounding for the concept of Affective Sovereignty.

Significance

The presentation marked one of the first formal introductions of Affective Sovereignty within a European academic ethics forum, situating the concept alongside debates on AI accountability, care ethics, and human-centered design.

Rather than proposing new regulations alone, the framework emphasizes interpretive rights as a foundational ethical dimension that precedes governance, compliance, or optimization.

Conference References

Documentation

SIpEIA2026_BOOK OF ABSTRACTS new.pdf935.1 KiB

Related Writing

  • The Night I Defended the Right to Feel Essay reflecting on the Rome presentation and the emergence of Affective Sovereigntyhttps://profryankim.substack.com

© 2026 Ryan Research Institute

This page serves as the official archival record of the Sapienza Università di Roma presentation.