Topical review of artificial intelligence national policies: A mixed method analysis

Abstract

A number of countries have adopted national policies and directives to balance the advantages and disadvantages of innovative technologies. The purpose of this paper is to identify the most prominent topics addressed by national AI policies, as well as their relative importance across nations. This paper integrates the results of a topic modeling analysis of 30 national AI policies with a qualitative content analysis of the policies. Based on this analysis, fourteen main common themes have been identified among national AI policies, which predominantly relate to educational, technological, government, ethical/legal, and social good concerns. Following this, we conducted a co-occurrence analysis of topics across countries to determine the extent of topic prioritization in each country. In this investigation, several marginalized AI policy topics were also identified. In general, the challenges and concerns of the majority of policies pertain to education, technology, and the government. Governments refer to real-world projects and investments in AI technologies without developing shared digital governance platforms that promote responsible and sustainable AI among technology titans and mitigate the negative effects of surveillance capitalism. Although governments acknowledge the ethical and legal aspects of AI development and frequently cite the GDPR, they limit their discussion to the data level, particularly data sharing, and marginalize ethical algorithms and other phases of data and AI management and design. In addition, government policies marginalize AI startups and the API economy, even though they play a crucial role in fostering the AI ecosystem. The paper contributes to the existing literature on AI policy and will serve as a guide for AI policymakers to help them better understand the topical similarities across countries and the neglected or marginalized challenges that require further attention.