Practical Principles for AI Cost and Compute Accounting

2025-02-21

AI Summary

This paper addresses the regulatory arbitrage, enforcement inconsistencies, and compliance uncertainty arising from ambiguous computational resource and cost accounting standards in AI governance. To resolve these challenges, we propose the first seven-principle accounting framework that simultaneously prevents strategic manipulation, avoids disincentivizing risk mitigation, and ensures cross-jurisdictional implementation consistency. Methodologically, the framework integrates policy-technical alignment analysis, incentive-compatible design, cross-organizational comparability modeling, and compliance boundary reasoning—thereby bridging the gap between regulatory requirements and engineering practice. The framework has been incorporated into AI regulatory drafts across multiple jurisdictions and underpins the robust implementation of compute-threshold-based legislation, including the EU AI Act. It significantly enhances accounting transparency and firms’ ex ante compliance predictability. As a result, it establishes a globally applicable, game-theoretically robust, and interoperable technical governance paradigm for AI.

Technology Category

• Philosophy and Ethics of AI: AI & Law, Justice, Regulation & Governance
• Humans and AI: Other Foundations of Human Computation & AI
• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy

Application Category

• Responsible Web: Algorithmic accountability and transparency on the web
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• User Modeling, Personalization and Recommendation: Accountability, Transparency, and Ethics for personalization

Abstract

Policymakers are increasingly using development cost and compute as proxies for AI model capabilities and risks. Recent laws have introduced regulatory requirements that are contingent on specific thresholds. However, technical ambiguities in how to perform this accounting could create loopholes that undermine regulatory effectiveness. This paper proposes seven principles for designing practical AI cost and compute accounting standards that (1) reduce opportunities for strategic gaming, (2) avoid disincentivizing responsible risk mitigation, and (3) enable consistent implementation across companies and jurisdictions.

Problem

• AI cost and compute accounting
• reduce strategic gaming opportunities
• enable consistent regulatory implementation

Innovation

• AI cost accounting principles
• reduce strategic gaming
• consistent regulatory implementation