Faculty Members React to ChatGPT and AI Software

On Nov. 30, 2022, OpenAI announced their new product ChatGPT which sparked debate among students, educators, parents and the New York Times about the future of education. ChatGPT is a novel chatbot trained to interact with users and respond to their questions. 

According to its developers, ChatGPT is trained to request more information when unclear, admit shortcomings in answers, correct biased premises and reject inappropriate requests. The tool is incredibly versatile and can be used to generate poetry, song lyrics, short essays and summarize information, among many others. In an academic context, some colleges are worried about its potential to be used to write essays and papers.  

However, the chatbot has a number of shortcomings. According to the developers, ChatGPT sometimes writes misleading or incorrect responses, it may overuse phrases or be excessively wordy and it may answer ambiguous questions incorrectly. One of the most discussed issues is the potential for the chatbot to produce biased or harmful content. According to Bloomberg News, this is because AI, like ChatGPT, relies on statistical association among words and phrases, which can result in the reproduction of biased ideas. To combat this issue, the developers are using Moderation API to restrict certain requests of the program. 

According to their mission, OpenAI is an artificial intelligence research and development company aiming to “ensure that artificial general intelligence (AGI)—by which we mean highly autonomous systems that outperform humans at most economically valuable work—benefits all of humanity.” Previously, they have put out products like DALL-E2, an AI that can create images and art from descriptions by users, and Codex, a system that converts written word into code. 

This technological development has not been missed by Bates faculty. The faculty met on Jan. 23 for a preliminary meeting to hear concerns and questions from professors about ChatGPT. Associate Dean of Faculty Donald Dearborn said that faculty are just beginning to “grapple individually with what this means for how they teach and how they do their work.”

According to Dearborn, faculty reaction is mixed. While there is concern about academic integrity, many are optimistic about its potential for use in the classroom. 

Professor of Psychology Todd Kahan has found ChatGPT useful as a preliminary tool to draft javascript code. “The code it generated needed some editing, but it helped me get started and if students encounter similar roadblocks in my lab-based classes, I wouldn’t have problems if they wanted to use this technology to generate computer code,” Kahan said.

That said, he’s found that “while [ChatGPT] can generate grammatically correct prose, the content is often repetitive and uninteresting. It’s possible that this aspect of ChatGPT will improve over time, but this will only make it more important to critically evaluate the content generated, as it may not always be accurate”

This finding was echoed by Professor of Philosophy Susan Stark, whose personal experiments with ChatGPT have returned basic discussions of topics that are oversimplified and contain “very little meaningful analysis.” 

“Like a computer or a calculator, ChatGPT can be used as a tool to overcome writer’s block, to help brainstorm ideas or approaches to a project,” Stark said. “If it is used, it might be used early in a project, where the student then takes up the bot’s suggestions and decides what to do with them in their own writing, speaking, and thinking.”

Dean Dearborn doesn’t expect that Bates will be “prescriptive institutionally” about the use of ChatGPT in the classroom. Instead, he anticipates a wide range of approaches. 

“At one end, there’s prohibiting the use of it and developing or using tools to detect the use of it like any other plagiarism or academic integrity issue,” Dearborn said. “At the other end [there is] developing instructive ways to use ChatGPT as a learning tool.” 

Regardless of approach, Stark believes substituting AI will not give students the benefits gained while learning to write. 

“An important part of both high school education, and a Bates education, is to promote students learning to read empathetically and critically, write and speak clearly, persuasively, and effectively, and think well,” Stark said. “A chat bot cannot think for anyone else, it is not even clear that the bots…