Algorithmic Orthodoxy and Gendered Scripture: AI, Biblical Interpretation, and Epistemic Injustice in the Global South

Abstract

This paper offers a critical, interdisciplinary examination of artificial intelligence ( AI ) as a theological mediator, arguing that AI systems do not merely transmit religious knowledge but actively reproduce gendered epistemic hierarchies embedded in their training data. Grounded in an African Technocultural Feminist framework and informed by feminist hermeneutics and critical AI studies, the study employs a qualitative, discourse-analytical method based on simulated interviews with major large language models. These prompts focus on women’s roles in church leadership, religious life in the Global South, and interpretations of Proverbs 31 and Ephesians 5:22–24, serving as test cases for identifying recurring algorithmic biases. The findings reveal a pattern described as “soft-neutral patriarchy,” where AI systems present balanced perspectives while subtly privileging dominant, often complementarian interpretations. Women are consistently represented in relation to domesticity, moral responsibility, and relational spirituality, while leadership roles are framed as conditional or exceptional. In interpretations of Proverbs 31, women’s agency is acknowledged but confined within domesticated and productivity-centered ideals. Similarly, readings of Ephesians 5:22–24 tend to soften hierarchical submission rhetorically while maintaining its structural logic, a tendency identified as “algorithmic equilibrium bias.” The study also shows how AI systems essentialize the Global South as inherently traditional, reinforcing colonial epistemic hierarchies and obscuring the diversity of African women’s lived experiences. Beyond biblical texts, the paper demonstrates how AI reproduces longstanding patriarchal distortions in representations of figures such as Mary Magdalene and Lilith, privileging sexualized or demonized portrayals over feminist reinterpretations. These dynamics are conceptualized as forms of testimonial injustice, hermeneutical erasure, and epistemic fragmentation, resulting in “algorithmic orthodoxy.” The paper concludes by advocating a decolonial reconfiguration of AI grounded in epistemic justice, indigenous knowledge systems, and participatory feminist design, positioning AI as a potential tool for theological liberation and gender justice in the Global South.