AI's Historical Predictions Reveal Systemic Biases in Data and Narratives
Original framing: “Can AI Be Trained to Predict the Past?” — Bloomberg
AI models often exclude or misrepresent marginalized voices in historical datasets, reinforcing dominant narratives. This omission perpetuates historical erasure and limits the diversity of perspectives in AI-generated historical analysis.
AI's attempt to predict the past is not a neutral exercise but a reflection of the dominant historical narratives encoded in its training data.
By excluding indigenous and marginalized perspectives, these models reinforce colonial and Eurocentric frameworks. A systemic approach requires integrating diverse historical sources, ethical auditing, and collaborative storytelling to ensure AI reflects a more inclusive and accurate understanding of history. This synthesis draws on cross-cultural perspectives, indigenous knowledge, and critical historical analysis to challenge the power structures that shape both AI and historical memory.