How VADER is your AI? Towards a definition of artificial intelligence systems appropriate for regulation
Leonardo C. T. Bezerra, Alexander E. I. Brownlee, Luana Ferraz Alvarenga, Renan Cipriano Moioli, Thais Vasconcelos Batista
TL;DR
The paper tackles the problem that AI regulation definitions often misclassify non-AI ICT and non-ICT works, risking overregulation. It introduces the VADER framework to score AI definitions against premises inspired by GDPR-like data privacy regulation and deliberate scope delimitation, supported by a representative dataset of techniques and examples. Using OECD 2019/2023 definitions as a grounding case and auditing major proposals (US AAA, EU AI Act, UK AI Bill, Brazil PL 2338), the authors find that none of the definitions are fully appropriate for regulation, with several posing concrete overregulation risks. The work emphasizes the need for precise AI-technique enumerations and international collaboration, and it provides an online repository for ongoing refinement of AI regulation definitions and scope.
Abstract
Artificial intelligence (AI) has driven many information and communication technology (ICT) breakthroughs. Nonetheless, the scope of ICT systems has expanded far beyond AI since the Turing test proposal. Critically, recent AI regulation proposals adopt AI definitions affecting ICT techniques, approaches, and systems that are not AI. In some cases, even works from mathematics, statistics, and engineering would be affected. Worryingly, AI misdefinitions are observed from Western societies to the Global South. In this paper, we propose a framework to score how validated as appropriately-defined for regulation (VADER) an AI definition is. Our online, publicly-available VADER framework scores the coverage of premises that should underlie AI definitions for regulation, which aim to (i) reproduce principles observed in other successful technology regulations, and (ii) include all AI techniques and approaches while excluding non-AI works. Regarding the latter, our score is based on a dataset of representative AI, non-AI ICT, and non-ICT examples. We demonstrate our contribution by reviewing the AI regulation proposals of key players, namely the United States, United Kingdom, European Union, and Brazil. Importantly, none of the proposals assessed achieve the appropriateness score, ranging from a revision need to a concrete risk to ICT systems and works from other fields.