Spatial_DataSet_for_Global public risk perception of artificial intelligence

This study conducted a thorough examination of public perceptions of artificial intelligence (AI) risks and their spatiotemporal development patterns. Employing a combination of Gensim, Perspective, and other natural language processing tools, alongside manual review methods, the data underwent systematic processing and categorization. Building on this foundation, the multidimensional mathematical convex hull approach was utilized to extrapolate public geospatial data, offering an in-depth analysis of the spatiotemporal distribution characteristics, regional heterogeneity, and potential influencing factors of global public perceptions towards AI risks. To mitigate the uncertainties inherent in methodological, data, and mathematical assumptions, the study integrated various machine learning models with the game theory-based Shap model to attribute and dissect the public's risk perceptions across different nations and regions, uncovering the impact of economic, political, and religious domains on public AI perception levels. Furthermore, the study elucidates the evolution of public risk perceptions towards AI over the past few decades. Ultimately, the empirical findings provided both theoretical and empirical bases for managers and policymakers to devise effective intervention strategies and social policies, aiming to foster a smooth and orderly development of AI technology, while advocating for the necessity of enhancing public risk awareness.