IBI-002354 - Severepublie
Knowledge Distillation Bias
Biais de résumé de connaissances
IBI Score
7.5
out of 10
Severity4_grave
Frequencyfrequent
Scope groups4
Detection difficulty4
🇫🇷 Définition (FR)
Biais transmis et amplifiés lors de la distillation d'un grand modèle vers un petit modèle.
🇬🇧 Definition (EN)
Bias transmitted and amplified when distilling a large model into a smaller one.
TriAxis Taxonomy
Structural Classification
O
Origin
O2
Computational & Statistical
L
Lifecycle
L5
Learning
S
Specific
S27
Emergent bias
💡
Concrete Example
Un petit modèle distillé à partir de GPT-4 reproduit et amplifie les biais du modèle source.
📊 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
- •
🛡 Impact & Mitigation
🎯 Vulnerable groups impacted
No specific groups.
🛠 Mitigation methodologies
- AUDIT_TIERS : Third-party Audit
🏢 Sectors concerned
Media & Social Platforms (Risk: /5)
⚖ Applicable regulations
🇺🇸 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
🇪🇺 Europe
- •AI_ACT_15 : AI Act (UE) 2024/1689
- •AI_ACT_9 : AI Act (UE) 2024/1689
🌐 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
a352a49a-6d7b-49e0-aac0-7a653f18c292
Explore other biases
250 documented biases, classified by the TriAxis taxonomy.