The AI Pentad, the CHARME$^{2}$D Model, and an Assessment of Current-State AI Regulation

2025-03-08

AI Summary

Current AI governance lacks a unified, actionable framework; prevailing value-oriented approaches—such as fairness and transparency—are hindered by conceptual subjectivity and poor operationalizability. Method: This paper proposes an ontology-based, componentized governance paradigm grounded in the “AI Pentad”—human/organizational actors, algorithms, data, compute, and energy—and introduces the novel CHARME²D model, enabling bidirectional mapping between these five core components and three regulatory enablers: registration/disclosure, monitoring, and enforcement. Leveraging cross-jurisdictional policy analysis and a structured evaluation matrix, the study conducts a comparative assessment of AI regulatory practices across the EU, China, the UAE, the UK, and the US. Contribution/Results: It identifies critical gaps—including insufficient transparency coverage, absence of compute and energy consumption regulation, and weakened organizational accountability—and establishes the first empirically grounded, assessable, and scalable systemic governance benchmark for global AI legislation, transcending the limitations of value-centric frameworks.

Technology Category

• Philosophy and Ethics of AI: AI & Law, Justice, Regulation & Governance
• Humans and AI: AI for Accessibility
• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy

Application Category

• Responsible Web: Technology governance, policy, and regulations
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications

Abstract

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 value-based 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$^{2}$D Model to explore further the relationship between the AI Pentad and AI regulatory enablers. Finally, we apply the CHARME$^{2}$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.

Problem

• Lack of unified approach to AI regulation
• Challenges in defining value-based AI regulatory methods
• Need for comparative assessment of AI regulatory frameworks

Innovation

• AI Pentad: five essential AI components
• CHARME2D Model: links AI components and regulation
• Comparative assessment of AI regulations globally