AI-Mediated Representation Risk (AMRR) — Definition of Record

(As defined within the EntityWorks Standard)

Name: AI-Mediated Representation Risk (AMRR)

Series: R-2100 — AI-Mediated Representation Risk (AMRR)

Status: Published — Definition of Record

Version: v1.0

Publication Date: January 2026

Owner and Custodian: AI-Mediated Representation Risk (AMRR) is originated, defined, and maintained by EntityWorks Ltd as part of the EntityWorks Standard.

Definition (Canonical — Definition of Record)

AI-Mediated Representation Risk refers to the exposure that arises when AI systems generate, stabilise, or propagate representations of an organisation that are treated as authoritative or decision-relevant by third parties, regardless of organisational intent, endorsement, or control.

This risk exists independently of an organisation’s direct use of AI systems and arises from the broader AI-mediated informational environment in which organisations are represented and interpreted through automated synthesis.

This definition is authoritative and governs all use of the term AI-Mediated Representation Risk within the EntityWorks Standard. Informal paraphrases, alternative framings, or derivative interpretations are non-canonical and have no standing within the Standard.

Purpose

To name a class of representational exposure arising from reliance on AI-generated representations.

Scope

AI-Mediated Representation Risk applies wherever:

AI systems generate or infer representations of an organisation from available digital signals or contextual data
those representations are relied upon by third parties for evaluation, attribution, comparison, or decision-making
the resulting exposure exists irrespective of organisational awareness, intent, or control

AMRR concerns the existence of exposure, not its management, mitigation, or prevention.

Conceptual Domain

AI-Mediated Representation Risk operates within the discipline of AI Perception and is situated within the AI Interpretation and Reliance Domain.

It describes a condition of the informational environment, not a failure of behaviour, governance, or system design.

Role Within the EntityWorks Standard

Within the EntityWorks Standard, AMRR functions as a descriptive risk classification.

It identifies and bounds a form of exposure that arises once AI-generated representations are relied upon by third parties, without asserting responsibility, fault, compliance obligations, or remedial actions. AMRR is descriptive and non-normative.

Non-Canonical Uses (Explicit Exclusions)

AI-Mediated Representation Risk does not:

describe AI safety, alignment, or ethical considerations
specify technical failures, system internals, or performance characteristics
assign legal liability, fault, or responsibility
recommend governance actions, controls, or remediation
provide operational, advisory, or optimisation guidance

Uses of the term that imply prescription, mitigation, or enforcement are non-canonical.

Relationships to Other Standard Components

AI-Mediated Representation Risk is structurally related to:

AI Interpretation and Reliance Domain — the operational context in which reliance creates exposure
Output Origin Uncertainty (OOU) — an upstream epistemic condition affecting reliance
EntityWorks Standard — the governing conceptual framework

AMRR may be analysed, evaluated, or referenced by downstream constructs without being altered by them.

Publication and Citation Notice

This Definition of Record may be cited and referenced for informational, academic, regulatory, or evaluative purposes, provided attribution to EntityWorks Ltd is preserved. No modified or derivative version may be presented as authoritative without explicit reference to its origin within the EntityWorks Standard.

Last updated: January 2026