Ava Wright discusses different ways professors are attempting to welcome AI use into the classroom | Ava Uhlack | The Poly Post

Ava Wright examines writing with ChatGPT: ‘CRAP’ in the classroom

By Ava Uhlack, May 7, 2024

Ava Wright, an assistant professor of philosophy from Cal Poly San Luis Obispo’s College of Liberal Arts, hosted a talk April 29 to discuss her upcoming paper, “On Crap and Writing with ChatGPT.”

In her paper, Wright argues that ChatGPT and other forms of Artificial Intelligence writing are threats to proper communication practice.

“AI writers produce a form of speech that I will refer to as ‘crap,’ and crap is antithetical to good communication,” Wright said as she opened her talk. Wright opened the talk defining the term “crap” as something of extremely poor quality and expanded on its presence in the writing/philosophical field. She defined “crap” as a pure form of “bs,” meaning that what is produced is a statement no one is intending to believe but is communicated anyway. The word “crap” is also an acronym. The “C” represents currency, asking how current the information is or if there is a publication date. The “R” stands for reliable, how truthful the information is and whether it be trusted. The “A” represents authority, which asks who posted the information and if it is an expert within their field. Lastly, the “P” refers to purpose, asking why it was written and if it possibly has a bias.

“Artificial intelligence is also under a noncooperative condition,” said Austin Hsueh, a philosophy student at Cal Poly Pomona. “If you ask it something and it doesn’t know the answer, the next output is not going to be, ‘I don’t know.’ It’s going to be the answer. So, it tries to guess the answer. You’re asking it to put out crap, even if you don’t mean to.”

AI is also becoming more common within the classroom. Based off statistics from the 2023-2024 school year, 63 percent of teachers found students’ work containing traces of generative AI in their schoolwork. Which is a large 48 percent increase from the last year, according to an article by EducationWeek. Wright described the results of AI writers as vague, obscure, irrelevant, incoherent and implausible.

“ChatGPT has been trained to try to not say anything false, and crap can be incoherent and contradictory,” Wright said. “No one is intended to believe crap, so it doesn’t need to make any sense.”

To demonstrate, Wright brought up the example of “The Kettle Logic.” This rhetorical device is a way to show how one uses multiple arguments to defend a point, but in doing so creates arguments that contradict one another. The premise of “The Kettle Logic” revolves around a lawyer defending a client against charges of damaging a kettle. To refute, defending counsel argues the client returned the kettle without a crack, that the kettle was cracked when the client borrowed it and that the client never borrowed the kettle in the first place. All the viewpoints expressed at once contradict and discredit the entire defense. Wright concluded AI writing is the same in its results with each product lacking one collective thing: human thought and intent.

Wright found AI writing tools being algorithms means they don’t have any sort of human moderation; it’s trusted to follow its code.

“If you take ownership of the thinking and beliefs you are producing in conjunction with the AI, there might be a way of using them,” Wright said. “Then you really are sharing what you think with the people, and you’re not just vandalizing these communicative channels, and you’re not just cutting yourself off from others.”

The talk concluded with a Q&A where discussions continued with how AI writing tools are faulty due to how humans have designed them. Some professors have tried to embrace the use of AI writing in the classroom. However, this just continues the cycle of empty communication that holds no value since there is no human intent behind the result, but just empty code, according to Wright.

“AI is going to be a part of a lot of things we are going to be working with so it’s good to be sensitive to the fact that it also has its downsides,” said Peter Ross, a professor of philosophy at CPP. “There will be ways in which it’s not going to be able to process like a human being and then there will be people who aren’t adequately served. You can imagine if all you had to work with was BillyChat as your advisor for fall.”

While there are possible benefits that could be reaped from AI writing use, right now it’s best to keep the “crap” speech usage to games like poker when bluffing, according to Wright.

Feature image courtesy of Ava Uhlack

Verified by MonsterInsights