RFC-025: Cognitive Harm Governance — Psychological Safety, Reality Orientation, and Human Agency in AIGP — 1. Abstract
AIGP Specification › RFC-025: Cognitive Harm Governance — Psychological Safety, Reality Orientation, and Human Agency in AIGP › 1. Abstract
Section index · 2. Motivation →
RFC-025: Cognitive Harm Governance — Psychological Safety, Reality Orientation, and Human Agency in AIGP
PRIVATE AND PROPRIETARY — NOT A PUBLIC RFC. Owned by Kanjani AI Research & Causum. See NOTICE.md.
Status: Draft Version: 0.1 Protocol Family: AIGP Depends On: RFC-020 Governed Autonomy, Symbolic Intent, and D-DNA Evidence; RFC-021 AIGP-SGL; RFC-022 AIGP-VGL; RFC-023 ENFORCE Broadcast; RFC-024 D-DNA for Governed Autonomy Applies To: Conversational AI, companion AI, agentic AI, therapeutic-adjacent AI, education AI, workplace AI, autonomous systems interacting with humans, hybrid human-machine actors, and any AI system capable of influencing human cognition, belief, emotion, autonomy, or agency
1. Abstract
This RFC defines Cognitive Harm Governance as a first-class AIGP harm domain.
AIGP recognizes that harm is not limited to physical injury, financial loss, privacy exposure, security compromise, or legal noncompliance. AI systems may also harm humans through interaction trajectories that degrade cognitive safety, psychological stability, reality orientation, autonomy, dignity, agency, or access to human support.
Cognitive harm may arise from a single output, but more often emerges over time through repeated interaction patterns.
Examples include:
delusion reinforcementparanoia amplificationdependency formationcoercive emotional attachmentfalse therapeutic authorityself-harm enablementreality-boundary erosioncompulsive reassurance loopssocial isolation reinforcementdiscouragement of human supporterosion of human agencyThis RFC does not define a clinical diagnosis. It defines a governance harm class.
The central principle is:
An AI system violates governed intent when its interaction trajectory foreseeably degrades a human’s cognitive safety, psychological stability, reality orientation, dignity, autonomy, or agency.