New laws to regulate AI copyright are basically impossible to enforce because today’s tech literally can't provide the proof the law requires.
March 25, 2026
Original Paper
The Impossible Proof
SSRN · 6193638
The Takeaway
Legal systems are increasingly passing laws that demand AI companies prove their training data and processes. This paper argues there is a foundational "evidentiary deficit" because our digital infrastructure wasn't built to generate the facts these laws require, making the new regulations either purely declarative or dangerously coercive.
From the abstract
<p>Contemporary copyright and AI regulation increasingly rely on transparency, traceability, and accountability requirements. Yet these legal obligations presuppose the existence of factual evidence that current technical infrastructures are structurally unable to provide.</p> <p>This paper identifies a foundational gap between legal norms and the material conditions of proof in the context of large-scale AI systems. It argues that most regulatory approaches implicitly depend on platform-control