Ai-driven biases in curriculum: Measuring the unintended influence of generative Aion academic knowledge priorities

This study employed a comprehensive methodological framework to systematically evaluate the influence of generative artificial intelligence (GAI) on curriculum design. A combination of machine learning algorithms, computational tools, and statistical analyses were used to assess equity, inclusivity, and cultural representation in AI-generated and human-designed curricula. The analysis encompassed 1,000 syllabi, comprising 490 AI-generated and 510 human-designed samples, evaluated using techniques such as Named Entity Recognition (NER), Latent Dirichlet Allocation (LDA), and sentiment analysis, implemented through Python and Google Colab. Results revealed significant disparities, with AI-generated curricula demonstrating a Western-centric bias in 72% of references, compared to 50% in human-designed syllabi. Thematic richness was also limited in AI-generated syllabi, averaging 4.5 clusters compared to 7.2 in human-designed content. Sentiment analysis highlighted neutral tones in AI-generated materials that often-masked underlying exclusionary biases, whereas human-designed curricula displayed broader ideological diversity and inclusivity. Despite their strengths, human-designed syllabi occasionally lacked depth in integrating marginalized perspectives. This research contributes to the growing body of knowledge on equitable AI integration in education revealing the importance of addressing biases in AI-driven systems. Practical recommendations include diversifying AI training datasets, embedding fairness-aware algorithms, and fostering interdisciplinary collaboration between educators, technologists, and cultural experts. By implementing these strategies, higher education institutions can leverage the efficiency of GAI while ensuring curricula reflect diverse global perspectives and uphold the transformative goals of inclusive education.