Ideological Homogenization from Value Embedding

Classification

Domain

1. Discrimination & Toxicity

Subdomain

1.3 Unequal performance across groups

Causal entity

Human

Intent

Intentional

Timing

Pre-deployment

Subdomain definition: Accuracy and effectiveness of AI decisions and actions is dependent on group membership, where decisions in AI system design and biased training data lead to unequal outcomes, reduced benefits, increased effort, and alienation of users.

Real-world incidents in this subdomain

• 27 Feb 2026Washington State DOL's AI Phone System Reportedly Failed to Provide Spanish-Language Service to Callers Requesting Spanish
• 5 Dec 2025UK Facial Recognition System Reportedly Exhibits Higher False Positive Rates for Black and Asian Subjects
• 11 Oct 2024Infinite Campus AI-Driven Student Risk Model Leads to Cuts in Support for Nevada's Low-Income Schools
• 6 Oct 2024Police Use of Facial Recognition Software Causes Wrongful Arrests Without Defendant Knowledge
• 23 Jun 2024Department for Work and Pensions (DWP) Algorithm Wrongly Flags 200,000 for Housing Benefit Fraud
• 25 May 2024Facewatch Reported to Have Wrongfully Flagged Home Bargains Customer as Shoplifter

How other frameworks describe this risk

• Bias and discrimination (value embedding)

  Regulating under Uncertainty: Governance Options for Generative AI (G'sell2024)

• Impact on affected communities

  AI Risk Atlas (IBM2025)

• Unfair capability distribution

  A Closer Look at the Existing Risks of Generative AI: Mapping the Who, What, and How of Real-World Incidents (Li2025)

• Disparate Performance

  Trustworthy LLMs: A Survey and Guideline for Evaluating Large Language Models’ Alignment (Liu2024)

• Fairness

  Trustworthy LLMs: A Survey and Guideline for Evaluating Large Language Models’ Alignment (Liu2024)

• Fairness

  A Survey of Artificial Intelligence Challenges: Analyzing the Definitions, Relationships, and Evolutions (Saghiri2022)

• Service/benefit loss

  Sociotechnical Harms of Algorithmic Systems: Scoping a Taxonomy for Harm Reduction (Shelby2023)

• Quality-of-Service Harms

  Sociotechnical Harms of Algorithmic Systems: Scoping a Taxonomy for Harm Reduction (Shelby2023)

Other entries from Maham2023

• Risks from Unreliability Risk Category
• Discrimination and Stereotype Reproduction Risk Sub-Category · 1.1
• Discrimination and Stereotype Reproduction Additional evidence
• Discrimination and Stereotype Reproduction Additional evidence
• Misinformation and Privacy Violations Risk Sub-Category · 3.1
• Misinformation and Privacy Violations Additional evidence
• Misinformation and Privacy Violations Additional evidence
• Accidents Risk Sub-Category · 7.3
• Accidents Additional evidence
• Accidents Additional evidence
• Accidents Additional evidence
• Accidents Additional evidence