The AI Pentad, the CHARME²D Model, and an Assessment of Current-State AI Regulation

Artificial Intelligence (AI) has made remarkable progress in the past few years with AI-enabled applications beginning to permeate every aspect of our society. Despite the widespread consensus on the need to regulate AI, there remains a lack of a unified approach to framing, developing, and assessing AI regulations. Many of the existing methods take a valuebased approach, for example, accountability, fairness, free from bias, transparency, and trust. However, these methods often face challenges at the outset due to disagreements in academia over the subjective nature of these definitions. This paper aims to establish a unifying model for AI regulation from the perspective of core AI components. We first introduce the AI Pentad, which comprises the five essential components of AI: humans and organizations, algorithms, data, computing, and energy. We then review AI regulatory enablers, including AI registration and disclosure, AI monitoring, and AI enforcement mechanisms. Subsequently, we present the CHARME<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>D Model to explore further the relationship between the AI Pentad and AI regulatory enablers. Finally, we apply the CHARME<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>D model to assess AI regulatory efforts in the European Union (EU), China, the United Arab Emirates (UAE), the United Kingdom (UK), and the United States (US), highlighting their strengths, weaknesses, and gaps. This comparative evaluation offers insights for future legislative work in the AI domain.