Data Quality and Bias

Real-world data used for training AI is often "dirty" (missing values, errors) and reflects societal biases. AI models learn these imperfections and biases.