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.
View on arXiv@article{casper2025_2502.15873,
title={ Practical Principles for AI Cost and Compute Accounting },
author={ Stephen Casper and Luke Bailey and Tim Schreier },
journal={arXiv preprint arXiv:2502.15873},
year={ 2025 }
}