Artificial intelligence (AI) governance is anticipated to have a transformative impact on humanity which has prompted researchers to analyze its implementation and use to ensure that the technology advances ethically and is beneficial for society. Though countries have begun to develop governance initiatives to regulate AI, the number of emerging AI regimes with an established structure is still relatively low. Meanwhile, the technology is advancing rapidly and has already caused harm inequitably to underrepresented communities. Thus, there is an urgent need to establish robust governance to mitigate the issues and risks attendant when deploying AI.While numerous ethics, principles, and structures have been recommended, this article intends to address the policy lag by providing policymakers with a simple and compelling AI governance framework that situates AI principles as the guiding baseline for developing and evaluating policies. Rather than devising new policy recommendations, the most recent (at the time of writing) and comprehensive governance documents from China, the European Union, and the United States were systematically selected, and examined in a comparative analysis to study how the three regimes address AI principles. Based on the comparative analysis, the most comprehensive and effective recommendations were selected to produce seven broad policy recommendations. The governance framework and recommendations are intentionally broad so that they can be adapted to adequately address AI principles across diverse contexts, encouraging the implementation of AI principles, increasing the likelihood of beneficial AI, and reducing the risks and harms associated with the technology. Nevertheless, the recommendations provided should not be considered exhaustive as the technology has an immense reach and new AI governance initiatives are developing continuously in this growth period in AI governance. It is thus essential for policymakers to survey the most current and relevant governance landscape to identify the best practices that are suitable for their specific context and need.
Ren Bin Lee Dixon, University of Minnesota Twin Cities
Policy Group Researcher at Center for AI and Digital Policy.
Журнал “AI and Ethics”, 2023.