A.I. Essays: Fraud in the Classroom or Final Frontier of Authorship?
AI writing tools like OpenAI’s GPT-3 and Google’s Lambda are transforming essay writing by generating coherent, human-like text in seconds, raising concerns about plagiarism and the integrity of academic work. While current copyright law does not protect AI-created works, debates around intellectual labor, creativity, and machine-driven outputs are intensifying. Critics worry that AI enables users to bypass the intellectual effort required in writing, while supporters argue it can make education more accessible and level the playing field for students with fewer resources. As AI evolves, it challenges institutions to rethink the boundaries of creativity and ownership.
Attribution
© Markus Winkler, Free to Use, via Pexels
Open AI’s GPT-3, Google’s Lambda, and The Good AI are writing essays in seconds. Initially, concerns over text-generating software were quelled when earlier programs produced unintelligible sentences and paragraphs that clearly did not resemble human writing. But automated writing is now coherent. With the advent of artificial intelligence writers, it is now possible for anyone with a computer to generate essays that are virtually indistinguishable from those written by humans. The process is also fast, greatly surpassing the standard time it takes to write a paper, making it easy for students to potentially plagiarize essays and difficult for professors to detect when an essay has been plagiarized. However, extending copyright protection to A.I. writing can fend off accusations of plagiarism and potentially recognize machine-driven creativity.
Copyright law today does not protect works created by artificial intelligence. In recent applications, the United States Copyright Office, which administers the national copyright system and provides advice on copyright law to Congress, federal agencies, the courts, and the public, has refused to extend protection and register creative works created by A.I. This is because copyright law explicitly states that a human creates a work to be eligible for protection.
One argument against A.I. writing tools is that they will help students cheat in the writing process with a tool that removes intellectual labor, whereas human authorship preserves integrity because it demands that students demonstrate an effort to complete an essay. Thus, artificial intelligence cheats the minimal degree of human labor involved in creativity, where one’s own labor should entitle a person to claim their intellectual fruits. If true, that principle would be at odds with the Supreme Court holding in Feist Publications, Inc. v. Rural Telephone Service Co., where the Court reasoned that the primary object of copyright is “[t]o promote the Progress of Science and useful Arts,” not “to reward the labor of authors.”
I spoke to Robert Kasunic, Associate Register of Copyrights and Director of Registration Policy and Practice for the Copyright Office, about intellectual labor and the human authorship requirement, and he said that the issue is not the presence of human labor. Rather, it is a question of finding a historical analogy and dealing with human input. For example, when using a camera, choices about the position of angles, the distance of the shot, the use of a particular lens, or deciding to enhance or subdue certain lighting or color would be examples of creative choices. Writing, a quintessential category of copyright, illustrates creativity when one crafts sentences, diction, tone, or character. Auto-generating software, according to the Copyright Office, removes even the most minimal of human decisions. In other words, typing a “command phrase” into a word processor, and having the computer generate that information back in logical form, skips choices that have been considered part of the creative process according to historically analogous tools.
A secondary concern is that artificial intelligence relies on other past works to train and create, some likely having copyright protections of their own. Thus, the use of A.I. creations can infringe on tens or hundreds of works at a time to arrive at its own creation. This could permit anyone to use or distribute A.I. creations without the copyright holder’s explicit permission. Therefore, without extending protection to A.I. works, there could be a chilling effect on the development of A.I. creations.
Courts should respond to the pace of these technological improvements while considering both the legal and commercial implications that will result from the deployment of A.I. in the classroom and beyond. As a tool, A.I. assistance can break down ableism in writing and make the process more accessible to people with difficulty with reading and writing, which is essential to persons with mobility, reading, and/or intellectual disabilities. An argument for minimal human creativity can be the selection of sentences for different tones or the arrangement of sentences in patterns that are personally conducive to flow. Institutions that derive revenue from academic paper submissions, essay exams, and writing preparation courses will also feel the impact of machine-generated essays if A.I. tools are able to shatter the glass ceiling of advanced writing. Currently, essay writing and exams are time-consuming and highly subjective, where even the topics chosen by students from different income households tend to generate unconscious bias during the application process. Students from more privileged backgrounds often have an advantage in essay writing and exams due to their access to better resources and tutors. A.I. writing systems could create a more impartial system of assessment and promote equal access to high-quality education regardless of background.
Originally Written on October 28, 2022