Beyond the False Dichotomy: Regulating AI Safety, Ethics, and Innovation
This paper was originally published in the Arizona State Law Journal.
Abstract
Artificial Intelligence (“AI”) is now at the center of regulatory attention amid debates over the European AI Act and proposals for an AI moratorium. Yet discussion is still framed by a false choice: regulators must supposedly pick between AI safety and ethics or promoting innovation.
This Article demonstrates the limits of a segregated approach. It challenges this oversimplification by dissecting two central dichotomies: first, between regulation for contemporary versus prospective AI, and second, between innovation and safety. By analyzing the regulatory impacts of the EU AI Act and a proposed U.S. regulatory moratorium in the contexts of non-discrimination and liability law, the Article shows that both extremes—comprehensive value-driven regulation and hands-off innovation policies—ultimately fail to achieve their intended aims and, in some respects, undermine their own normative foundations.
Neither sweeping ethical mandates nor regulatory inaction meaningfully address the practical and legal challenges created by AI’s opacity, unpredictability, and complexity. Instead, both approaches risk overinclusive or underinclusive frameworks that fragment legal protections and reduce certainty for all stakeholders.
Rejecting the idea that effective AI regulation requires a trade-off, the Article proposes a Pareto principle for AI governance: narrow, targeted procedural safeguards—such as robust documentation and clear accountability mechanisms—can maximize AI safety with minimal cost to innovation. Rather than reinventing the regulatory wheel, these measures should be anchored in established legal regimes, equipping regulators to address both contemporary and prospective AI developments. By moving beyond the false dichotomy, this Article offers a roadmap for sustainable, high-impact AI regulation that preserves both innovation and public trust.