arXiv:2606.07629v1 Announce Type: new 
Abstract: Current approaches to aligning large language models (LLMs) aggregate diverse human preferences into a single reward signal, effectively optimizing for a hypothetical ``average user'' who represents no real person particularly well. This position paper argues that LLMs should learn personalized, individual preferences rather than aggregated ones. We show that aggregation masks critical information about preference diversity, individual values, and contextual dependencies, which is a limitation both theoretically grounded in social choice theory and empirically evident across demographic groups. We analyze the rich structure that human preferences encode, survey technical approaches to personalization, and systematically address counterarguments on scalability, shared standards, and manipulation risk. While personalization introduces genuine safety challenges including filter bubbles, value lock-in, and psychological manipulation, we argue these are manageable through bounded personalization frameworks that preserve universal safety constraints while accommodating legitimate individual variation. We conclude with a concrete research and policy agenda for developing preference-aware models that respect both individual autonomy and collective safety.

تشير ورقة موقف حديثة إلى أن نماذج اللغة الكبيرة (LLMs) يجب أن تعطي الأولوية لتعلم التفضيلات البشرية الشخصية بدلاً من المجمعة، حيث إن الأساليب الحالية تعمل على تحسين تجربة مستخدم متوسط افتراضي، مما لا يمثل الأفراد الحقيقيين بشكل فعال. يسلط هذا النهج الضوء على أهمية التعرف على تنوع التفضيلات والاعتماديات السياقية.

Un reciente documento de posición argumenta que los modelos de lenguaje de gran tamaño (LLMs) deberían priorizar el aprendizaje de preferencias humanas personalizadas en lugar de agregadas, ya que los métodos actuales optimizan para un usuario promedio hipotético, que no representa de manera efectiva a personas reales. Este enfoque destaca la importancia de reconocer la diversidad de preferencias y las dependencias contextuales.

Un récent article de position soutient que les modèles de langage de grande taille (LLMs) devraient privilégier l'apprentissage des préférences humaines personnalisées plutôt que celles agrégées, car les méthodes actuelles optimisent un utilisateur moyen hypothétique, qui ne représente pas efficacement les véritables individus. Cette approche souligne l'importance de reconnaître la diversité des préférences et les dépendances contextuelles.

A recent position paper argues that large language models (LLMs) should prioritize learning personalized human preferences over aggregated ones, as current methods optimize for a hypothetical average user, which fails to represent real individuals effectively. This approach highlights the importance of recognizing preference diversity and contextual dependencies.