Dataset for: The Cognitive Overload Paradox: An LLM Powered Systematic Review of Affective-Cognitive Ecology in Human-AI Interaction

This dataset contains the extracted features from 806 peer-reviewed empirical studies (2016–2026) regarding affective-cognitive ecology in AI-assisted language learning. The extraction employed a Human-in-the-Loop (HITL) framework using GPT-4o. This version includes refined data, where cross-disciplinary outliers (e.g., geosciences, cloud computing) were identified and removed. This dataset supports the findings presented in the article