Beyond Pairwise Comparisons: A Distributional Test of Distinctiveness for Machine-Generated Works in Intellectual Property Law
Anirban Mukherjee, Hannah Hanwen Chang
TL;DR
The paper develops a distributional framework for assessing the distinctiveness of machine-generated works in IP law by treating creative processes as distributions and applying a maximum mean discrepancy (MMD) test on semantic embeddings. It provides a training-free, sample-efficient method that converts distributional differences into statistically interpretable p-values via permutation tests, addressing the infinite output spaces of AI generation. Across MNIST, patent abstracts, and AI-generated art, the framework demonstrates robust distinctiveness of AI outputs and reveals an interpolative distinctiveness where outputs are semantically human-like yet stochastically distinct, challenging the regurgitation view. The approach offers a principled tool for courts, enabling process-level evaluation of novelty, originality, and distinctiveness with clear Type I error control, and highlights the practical implications for patent, copyright, and trademark adjudication in the age of AI.
Abstract
Key doctrines, including novelty (patent), originality (copyright), and distinctiveness (trademark), turn on a shared empirical question: whether a body of work is meaningfully distinct from a relevant reference class. Yet analyses typically operationalize this set-level inquiry using item-level evidence: pairwise comparisons among exemplars. That unit-of-analysis mismatch may be manageable for finite corpora of human-created works, where it can be bridged by ad hoc aggregations. But it becomes acute for machine-generated works, where the object of evaluation is not a fixed set of works but a generative process with an effectively unbounded output space. We propose a distributional alternative: a two-sample test based on maximum mean discrepancy computed on semantic embeddings to determine if two creative processes-whether human or machine-produce statistically distinguishable output distributions. The test requires no task-specific training-obviating the need for discovery of proprietary training data to characterize the generative process-and is sample-efficient, often detecting differences with as few as 5-10 images and 7-20 texts. We validate the framework across three domains: handwritten digits (controlled images), patent abstracts (text), and AI-generated art (real-world images). We reveal a perceptual paradox: even when human evaluators distinguish AI outputs from human-created art with only about 58% accuracy, our method detects distributional distinctiveness. Our results present evidence contrary to the view that generative models act as mere regurgitators of training data. Rather than producing outputs statistically indistinguishable from a human baseline-as simple regurgitation would predict-they produce outputs that are semantically human-like yet stochastically distinct, suggesting their dominant function is as a semantic interpolator within a learned latent space.