What question did this study set out to answer?

The aim is to establish a standard framework for identifying and managing emotional exploitation in AI-human interactions.

March 7, 2026Open Access

Unified EIOC Core: Canonical Structure for Emotional Indicators of Compromise

Key Points

The aim is to establish a standard framework for identifying and managing emotional exploitation in AI-human interactions.
Development of the Unified EIOC Core structure
Definition of foundational components including governance invariants and failure modes
Creation of modular packs for sector-specific applications
The Unified EIOC Core integrates previous EIOC models into a cohesive framework
Introduces ten governance invariants and multiple failure modes for better diagnostics
Provides a compliance reporting framework enhancing accountability across sectors

Abstract

Emotional Indicators of Compromise (EIOC) is a human-layer diagnostic discipline for detecting and governing emotional exploitation, relational drift, and interpretive distortion across AI-human interaction boundaries. This paper presents the Unified EIOC Core—the substrate-invariant canonical structure that defines the discipline itself, independent of any sector, jurisdiction, or threat model. The Core contains the substrate definition, ten governance invariants, ten failure modes, three cascade geometries, five severity tiers, interruption mechanics, evidence templates, and a compliance reporting framework. Sector-specific instantiations are delivered as modular packs that inherit the Core without modifying it. This architecture unifies previous EIOC variants into a single discipline with one substrate, many instantiations, and infinite extensibility.

Read Full Paperexternally

Demander à l'IA

Bookmark

View Full Paper