The Age of Industrialized Plagiarism is Here

Photo by Ilya Pavlov on Unsplash

The explosion in the use and discussion of ChatGPT, Midjourney and similar tools has stormed across an ethical Rubicon, and the breaches that are occurring aren’t getting enough attention, in my opinion. If you’ve been living under a rock, a slew of so-called Artificial Intelligence (AI) tools have been brought online, and have captured the zeitgeist. The programs respond to prompts from users and produce responses that mimic human language and art. I preface the term AI with "so-called" because these machines do not think. For example, ChatGPT and its competitors are large language models that provide the statistically most likely responses to the questions they are posed. The results can seem impressive for a machine. The prose can even be witty and clever. The images Midjourney and similar programs create are notoriously terrible at drawing hands, but crop the grotesque appendages out, and the programs produce what looks to untrained eyes like high-quality, time-intensive "art." How are these machines at least somewhat convincingly aping human creativity?

The phrase "statistically most likely" is a key breadcrumb on the trail. Statistically most likely compared to what? Let’s phrase the question another way: If I ask ChatGPT to write me a speech in the oratory style of Barack Obama and it provides a close approximation, what were the references the program used? When Midjourney produces an interesting Avengers-related image, how does it know what the Avengers are and what they look like? What are the visual references? These programs are sitting on top of incomprehensibly enormous volumes of reference material, much of which is still under copyright, and none of which the creators agreed to have used and monetized in this way. Whenever these programs respond to queries, it is very likely that someone’s (probably multiple someones) work has been used without their permission to generate a response. These creators are also not credited. This is an enormous ethical breach, in my opinion, particularly when the programs facilitate copycats stealing from small, powerless creators.

Even more troubling, the chatbot programs have the capacity to be explicitly dishonest. They have been caught fabricating sources. It tracks, when you think about it. These machines are sitting atop huge repositories of pilfered work that they’ve…