William & Mary Bill of Rights Journal
Abstract
This Note argues (1) that the unique character of government copyright infringement supports a copyright holder’s use of the Takings Clause of the Fifth Amendment, as opposed to reliance on the Copyright Act for downstream Gen A.I. suits; (2) that Gen A.I. developers may wage similar claims once asserting authorship rights and copyright over their outputs; and (3) that the viability of federal government copyright infringement as a regulatory taking could change the way legal actors deal with copyrighted material. More modestly assessed, an answer on whether the Takings Clause protects copyrights would offer much needed guideposts as intellectual property rapidly changes hands and loses attribution on the internet. The pivotal impact of artificial intelligence on the world seems a welcome avenue for these questions. Any guidance from courts on copyright takings, the bounds of A.I. company rights, and where Gen A.I. fits into the future of intellectual property is helpful for all parties. This Note suggests that seeking answers to these questions would provide clarity—even if only slightly—to the litany of uncertainty ahead. These answers, however, could just as likely cut against A.I. companies as they could the U.S. government, sketching an interesting interplay between government power and corporate control.
Part I provides background on A.I. and the Takings Clause, foreshadowing a resurgence in intellectual takings claims. Part II analyzes copyright takings and asserts that the Takings Clause is a proper path for recharacterizing federal governmental copyright “infringement,” arguing that private individuals suing Gen A.I. developers for copyright infringement could sue the U.S. government under a constitutional takings claim but face an uphill battle under the Penn Central factors and general pleading standards. Part III suggests that Gen A.I. developers will likely assert authorship over copyright outputs, opening new takings questions, but also that these companies’ complicated and questionable machine learning procedures offer weak footing. Part IV navigates side effects and the practicability of this litigation, offering recommendations for Gen A.I. companies, copyright holders, and government officials.
This abstract has been taken from the author's introduction.