Against AI Exceptionalism: Traditional Product Liability Law as Sufficient Framework for AI Systems

Hanting Fan

Keywords: artificial intelligence, product liability, AI exceptionalism, tort law, design defect, failure to warn

Paper Type: Short papers / work-in-progress

TL;DR: Traditional product liability doctrine may provide a sufficient framework for AI accountability, questioning the necessity of AI-specific liability regimes.

Abstract: This paper challenges AI exceptionalism—the claim that neural network opacity renders traditional tort doctrine inadequate for artificial intelligence systems. Drawing on the landmark Benavides v. Tesla verdict of August 2025, where a Florida jury applied conventional design defect and failure-to-warn doctrines to award $243 million against Tesla's Autopilot system, this analysis demonstrates that existing product liability frameworks suffice for AI accountability. The paper examines technical challenges posed by deep learning systems and critiques emerging AI-specific regimes, including the EU's revised Product Liability Directive and the proposed U.S. AI LEAD Act. It concludes that traditional tort law has governed complex products with opaque mechanisms for decades, and the path forward requires rejecting AI exceptionalism rather than codifying it.

Submission Number: 46