IBI-000485 - Criticalpublie
Cascading Delegation Bias
Biais de délégation en cascade
IBI Score
9.0
out of 10
Severity5_critique
Frequencyfrequent
Scope groups5
Detection difficulty5
🇫🇷 Définition (FR)
Biais amplifiés lorsque les agents délèguent à d'autres agents sans vérification.
🇬🇧 Definition (EN)
Bias amplified when agents delegate to other agents without verification.
TriAxis Taxonomy
Structural Classification
O
Origin
O2
Computational & Statistical
L
Lifecycle
L7
Deployment
S
Specific
S02
Feedback loop bias
💡
Concrete Example
Un agent de recrutement délègue à un agent de scoring qui délègue à un agent de vérification, chacun ajoutant du biais.
📊 Detection Metrics
Feedback Loop RatioDemographic Parity Difference
⚙ Alert Thresholds
Feedback Loop Ratio0.15
Demographic Parity Difference0.1
📚 Sources & Scientific References
- •Multi-Agent Risks from Advanced AILewis Hammond, Alan Chan, Jesse Clifton, Jason Hoelscher-Obermaier, Akbir Khan et al. (2025)
- •Delegated Misalignment: How Multi-Agent Structures Amplify LLM Safety RisksZonghao Ying, Jiaqi Yan, Huize Luo, Quanchen Zou, Aishan Liu, Xianglong Liu (2026)
🛡 Impact & Mitigation
🎯 Vulnerable groups impacted
No specific groups.
🛠 Mitigation methodologies
- AUDIT_TIERS : Third-party Audit
🏢 Sectors concerned
Finance & Banking (Risk: /5)HR & Recruitment (Risk: /5)
⚖ Applicable regulations
🇪🇺 Europe
- •AI_ACT_14 : AI Act (UE) 2024/1689
🌍 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
📋 Metadata
Contributor
Mohamadou SIDIBE
Institution
AI BIAIS INITIATIVE
Version
1
Created on
October 14, 2026
Updated on
November 4, 2026
UUID
23703f0c-1404-45b9-a5c5-e9560cd2e6f2
Explore other biases
250 documented biases, classified by the TriAxis taxonomy.