UNFCCC：2024年发展中国家人工智能驱动的气候行动研究报告：机遇、挑战与风险（英文版）

AI is a multifaceted field with many plausible definitions, but definitions typically focus on an engineered system’s ability to perform some form of intelligent-seeming behaviour, such as generating content, forecasts, recommendations or decisions.1 Machine learning (ML), a subfield of AI, is concerned with computational techniques for enabling a model’s behaviour to reflect training data or experience.2 Climate-relevant subfields of AI include computer vision and image recognition, natural