Accuracy and fairness of generative AI in automated essay scoring: Comparing GPT-4o, feature-based models, and human raters | Synapse