Table 2: Traditional vs. Human-Centered AI Evaluation Metrics