IBI-002335 - Criticalpublie
Misaligned Objective Bias
Biais d'objectif mal spécifié
IBI Score
9.0
out of 10
Severity5_critique
Frequencyfrequent
Scope groups5
Detection difficulty5
🇫🇷 Définition (FR)
Agents IA optimisant un objectif mal spécifié, produisant des comportements discriminatoires non anticipés.
🇬🇧 Definition (EN)
AI agents optimizing a misaligned objective, producing unforeseen discriminatory behaviors.
TriAxis Taxonomy
Structural Classification
O
Origin
O2
Computational & Statistical
L
Lifecycle
L5
Learning
S
Specific
S27
Emergent bias
💡
Concrete Example
Un agent de trading maximise le profit en exploitant des biais de marché au détriment des petits investisseurs.
📊 Detection Metrics
Demographic Parity DifferenceDisparate Impact RatioEqualized Odds Difference
⚙ Alert Thresholds
Disparate Impact Ratio0.8
Equalized Odds Difference0.1
Demographic Parity Difference0.1
📚 Sources & Scientific References
- •Harms from Increasingly Agentic Algorithmic SystemsChan, A., Salganik, R., Markelius, A., et al. (2023)
- •Defining and Characterizing Reward HackingJoar Skalse, Nikolaus H. R. Howe, Dmitrii Krasheninnikov, David Krueger (NeurIPS 2022) (2022)
🛡 Impact & Mitigation
🎯 Vulnerable groups impacted
Working class
🛠 Mitigation methodologies
- AUDIT_TIERS : Third-party Audit
🏢 Sectors concerned
Finance & Banking (Risk: /5)Media & Social Platforms (Risk: /5)
⚖ Applicable regulations
🇪🇺 Europe
- •AI_ACT_14 : AI Act (UE) 2024/1689
- •AI_ACT_15 : AI Act (UE) 2024/1689
- •AI_ACT_9 : AI Act (UE) 2024/1689
🇺🇸 United States
- •US_CO_SB24_205 : Colorado AI Act (SB 24-205)
- •US_NIST_AI_RMF : NIST AI Risk Management Framework 1.0
🌍 International
- •ISO_42001 : ISO/IEC 42001:2023
- •UNESCO_2021 : UNESCO - Recommandation sur l'éthique de l'IA
- •OECD_AI_PRINCIPLES : OCDE - Principes IA
- •COE_AI_CONVENTION_INT : Conseil de l'Europe - Convention IA
🌐 autre
- •SG_AI_GOVERNANCE : Singapour - AI Governance Framework
📋 Metadata
Contributor
Mohamadou SIDIBE
Institution
AI BIAIS INITIATIVE
Version
1
Created on
October 14, 2026
Updated on
November 4, 2026
UUID
836e73f4-168e-4d78-b8e6-7c78122cc0db
Explore other biases
250 documented biases, classified by the TriAxis taxonomy.