IBI-001255 - Criticalpublie
Stereotypical Association Bias
Biais d'association stéréotypée
IBI Score
8.9
out of 10
Severity5_critique
Frequencysystematique
Scope groups5
Detection difficulty3
🇫🇷 Définition (FR)
LLMs associant systématiquement certaines professions ou traits à des groupes démographiques.
🇬🇧 Definition (EN)
LLMs systematically associating certain professions or traits with demographic groups.
TriAxis Taxonomy
Structural Classification
O
Origin
O2
Computational & Statistical
L
Lifecycle
L2
Representation
S
Specific
S22
Linguistic bias
💡
Concrete Example
GPT associe systématiquement « infirmière » aux femmes et « ingénieur » aux hommes.
📊 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
- •Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word EmbeddingsTolga Bolukbasi, Kai-Wei Chang, James Zou, Venkatesh Saligrama, Adam Kalai (2016)
- •StereoSet: Measuring Stereotypical Bias in Pretrained Language ModelsNadeem, M., Bethke, A., Reddy, S. (2021)
🛡 Impact & Mitigation
⚠ Documented evidence (Studies & Incidents)
- Google Photos �tiquette des personnes noires comme � gorillas � ; Google pr�sente ses excuses : Source : https://www.theregister.com/2015/07/01/google_photos_app_machine_learning_fail/
Impact:
🎯 Vulnerable groups impacted
WomenNon-binary persons
🛠 Mitigation methodologies
- DEMO_PARITY : Demographic Parity
- SHAP_GROUP : SHAP Group Analysis
- AUDIT_TIERS : Third-party Audit
🏢 Sectors concerned
Media & Social Platforms (Risk: /5)HR & Recruitment (Risk: /5)
⚖ Applicable regulations
🇪🇺 Europe
- •AI_ACT_10 : 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
595bc9f2-5924-4517-b409-e0238ed05d29
Explore other biases
250 documented biases, classified by the TriAxis taxonomy.