Preprint: “The Files are in the Computer: Copyright, Memorization, and Generative AI”

The essay (preprint) linked below was recently shared on arXiv.

Title

The Files are in the Computer: Copyright, Memorization, and Generative AI

Authors

A. Feder Cooper
Cornell University

James Grimmelmann
Cornell Tech and Cornell Law School

Source

via arXiv

DOI: arXiv:2404.12590
April 19, 2024

Abstract

A central issue in copyright lawsuits against generative-AI companies is the degree to which a generative-AI model does or does not “memorize” the data it was trained on. Unfortunately, the debate has been clouded by ambiguity over what “memorization” is, leading to legal debates in which participants often talk past one another. In this essay, we attempt to bring clarity to the conversation over memorization.

From the Essay

We take no position on what the most appropriate copyright regimes for generative-AI systems should be, and we express no opinion on how pending copyright lawsuits should be decided. Our goal is merely to describe how these systems work so that copyright scholars can develop their theories of generative AI on a firm technical foundation. We seek clarity, precision, and technical accuracy.

Access to Full Text Essay
39 pages; PDF.

UPDATE (April 26, 2024) Advocates Urge Law Journal to Disclose Microsoft, Google Ties (via Bloomberg)

Filed under: Data Files, News