The increasing integration of artificial intelligence (AI) in architecture and construction has revolutionised design processes. Yet, concerns persist regarding algorithmic biases reinforcing gender and cultural disparities in the built environment. This study critically examines how AI-driven systems reflect and perpetuate these biases through training datasets, algorithmic decision-making, and spatial design outcomes. Grounded in an Afrocentric perspective, this research challenges Eurocentric architectural norms and emphasises the need for inclusive, culturally sensitive AI applications. A qualitative methodology was employed, incorporating a systematic literature review, case study analysis, and theoretical synthesis using Value-Sensitive Design and algorithmic justice frameworks. The findings indicate that AI in architecture disproportionately privileges Western design principles, marginalises African architectural traditions, and exacerbates historical inequities, influencing access to spatial resources, urban planning decisions, and professional opportunities. To address these issues, the study advocates for fairness-driven AI models, diverse and representative training datasets, and participatory design approaches that integrate indigenous knowledge systems and liberatory design principles. In embedding Afrocentric and intersectional perspectives into AI-driven architectural tools, this research contributes to broader discussions on AI ethics, human-AI collaboration, and spatial justice while offering strategic recommendations for policymakers, designers, and technologists to ensure a more equitable and culturally inclusive built environment.