International Journal of Management

Algorithmic recommendation systems play a central role in shaping digital experiences across platforms such as YouTube, TikTok, Netflix, Spotify, Amazon, and Google. These systems use machine learning models trained predominantly on behavioral data to provide personalized content, products, and information. However, when applied globally, recommendations may reflect cultural biases embedded in training data, objective functions, user modeling assumptions, moderation policies, and linguistic coverage. This research analyzes sources of cultural bias in recommendation systems and demonstrates, through hypothetical multi-country datasets, how algorithms can reinforce cultural homogenization, suppress minority content, and misinterpret culturally distinct behavioral signals. A multi-layer cultural fairness framework is proposed to enhance inclusive algorithmic design.