Section: Arts

Comitta discusses authorship and artificial intelligence

What do some of the best books have in common? The worst? Can an entire first chapter be written solely using the first lines of other books? Tom Comitta covered these topics during their talk and reading titled “Human AI” hosted on Oct. 2 in which they read snippets of their novels: The Nature Book, People’s Choice Literature: The Most Wanted and Unwanted Novels and Patchwork

Assistant Professor of English Alyssa Quinn kicked off the event by thanking those who made it possible before handing the presentation over to Jonah Hyre ’27, who introduced Comitta and their books. “Reading Patchwork is like dreaming,” Hyre remarked. “Through the reassembly of fragments, once familiar things become new, strange and wonderful.” 

The first book Comitta discussed was People’s Choice Literature: The Most Wanted and Unwanted Novels. They began by referencing two similar studies that used polls to discover general preferences. The first focused on the paintings most and least wanted by people in the United States and the Soviet Union. The second turned its attention to the most and least wanted songs (the least wanted was 20 minutes long and included Walmart advertisements, prompting laughter from the crowd). 

Then, Comitta moved on to discuss the logistics of how they made their book. People’s Choice Literature was split into two parts. The Most Wanted was first, and The Least Wanted was second; however, either of the two sections could be read first. One tool they used was a large language model (LLM), but the program came out before ChatGPT and the generated responses were heavily selected and edited. For the more wanted novel, thriller and mystery were the most common genres. “In the most unwanted category,” Comitta countered, “we have romance, horror, historical fiction and classic literature.” The stark differences in genres and audience preference are also shown through the protagonists of the books. For The Most Wanted Novel, the protagonist Alix Finn is a straight white woman with good morals who falls in love. For The Least Wanted Novel, the protagonist Lord Jimothy Tickletext is a 160-year-old genetically-engineered billionaire trust-fund child on Mars. When Comitta began the reading of The Most Wanted Novel, the audience sat rapt and quiet. However, for The Least Wanted Novel, the audience gasped and chuckled as Comitta read the strange and experimental writing.  Both books entertained the audience, but the differing reactions were in clear contrast.

Next, Comitta presented Patchwork, a book made completely out of other books. Each chapter had its own theme. For example, the first chapter consisted of other books’ introductions, while another was made up entirely of sentences with typos and mistakes. “There’s an entire chapter told without words, just illustrations from Charles Dickens novels that I kind of collaged into a continuous image,” Comitta noted. They then read bits and pieces from throughout the book. The crowd particularly enjoyed a chapter formed with sentences using the word “ejaculate.”

Finally, Comitta discussed their first book: The Nature Book. Like Patchwork, this book was created entirely out of found language. Comitta then read a sound poem from the book, which consisted entirely of onomatopoeia found in other nature writing like wooshes, cracks and clicks.

Comitta’s writing emphasizes that there is a way to build off of others’ work without the use of a chatbot like ChatGPT. Comitta noted that, while they did utilize an LLM when creating People’s Choice Literature and Patchwork, AI was still unable to write stories from the prompts that they provided. The interest garnered from the absurdity of the most unwanted novel was not generated by ChatGPT but instead made and imagined by a human being. While the words of Patchwork and The Nature Book were taken from other books, the final products were wholly original.

0 Comments

Comments for this article have closed. If you'd like to send a letter to the editor for publication, please email us at collegian@kenyon.edu.