National AI Policy: Keyword and Topic Modelling Analysis

Résumé fourni par la source

This paper undertakes an Artificial Intelligence/Machine Learning based assessment of AI policies by analyzing national strategies and other relevant national documents on Artificial Intelligence. The paper has shortlisted 22 National AI policy documents from various countries, making it exploratory. These documents undertake two levels of computations: keyword analysis and topic modelling, thereby following a Computationally Grounded Theory (GCT) approach. The paper empirically establishes the user-provider trend towards shaping a sound AI policy. The paper also groups the countries with similar national AI policy focus through topic modelling. The novelty of the paper lies in three aspects. Firstly, the paper uses two independent analysis layers: keyword analysis and topic modelling. Secondly, the paper is neutral from author bias as the results are based on unsupervised algorithmic computations. Finally, the paper compares National AI strategy from a policy document standpoint based on similarities and differences. The paper establishes that, though GCT is effective in policy application, a qualitative framework cannot be done away with.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.