GPT Chatbots and Academic Integrity: Why Generative AI Should be Allowed in the Classroom

The International Baccalaureate, a prestigious college prep organization, recently allowed students to use ChatGPT on the IB exam as long as they do not claim it as their work. This surprising development is the latest in the string of headlines related to generative AI. Many school systems have already banned ChatGPT out of fear of academic dishonesty. Turnitin has announced the development of an AI-detection software that will be rolling out in April. Generative AI has made many rapid improvements, but students will not completely forego traditional learning in favor of bot-generated responses. Educators should integrate AI into the classroom so students know how to use the developing technology effectively and ethically.

Limitations with AI

An educator’s biggest fear is that students will use ChatGPT to do their work for them. While this is a valid fear it is often exagerrated. Olya Kudina’s students were tasked with comparing their assignments with an AI generated assignment. They were initially blown away by “how quickly the chatbot rendered information into fluid prose;” that is until they reread the bot’s work. The students realized that the bot was using incorrect information and it was unable to provide sources for the things it was claiming . Kudina’s students concluded that “copying from ChatGPT wouldn’t actually net them a good grade.”

On the other hand, Pieter Snepvangers received a passing grade on a 2,000 word essay that was written by AI. While AI generated work is passable at first glance, it fails to produce high quality academic work. Snepvangers’ essay was very generalized and didn’t include any citations. The lecturer grading the essay said it had “fishy language.”

Many teachers are horrified at the thought that students will be able to type an essay prompt into an AI generator and submit a paper in about 20 minutes; however it is not always that simple. ChatGPT can only produce up to 365 words, meaning that the student must ask the AI multiple questions to meet a higher word count. Snepvanger asked ChatGPT ten different questions related to the essay prompt. After receiving the paragraphs, Snepvanger selected the best ones and “copied them in an order that ‘resembled the structure of an essay,’”  Even when trying to use Chat GPT, students must use fundamental writing skills to create a coherent paper.

An AI is unlikely to do well on any assignment that requires anything more than a surface-level understanding of the material.  AI can’t properly source an article, and it doesn’t tell its users where its information came from. Any analysis produced by AI will be far too shallow to receive a good grade because the AI is unable to understand the deeper themes of texts or form original perspectives.

AI and Academic Integrity

Many institutions are grappling with the ethics of allowing AI in the classroom. Some are simply requiring that students state when they are using AI generated text. Others are claiming that use of AI in any form is plagiarism. Kalley Huang equates the current status of ChatGPT to Wikipedia in the early 2000’s, which educators also saw as the end of traditional education. Villanova University’s Chair of the Academic Integrity Program, Alice Dailey, believes that schools should allow AI, but they should develop a blanket policy that covers a wide range of circumstances. The advancement of AI technology is going to force educators revolutionize the way they evaluate student progress.

Many schools are moving away from take home essays and are using “in-class assignments, handwritten papers, group work and oral exams” to combat AI plagiarism. Stephen Marche claims that the essay is the “way we teach children how to research, think, and write.” Though these skills are important, there are better ways to teach them in an AI world. Antony Aumann is requiring students write their first draft in the classroom and explain each revision. This method of teaching is not only teaching the students how to write an essay, but is also forcing them to think critically about why they are writing it in this way.

AI in the Classroom

Instead of banning ChatGPT because of its risk to academic integrity, educators should use it as a tool in the classroom. The technology is not going away and students will be unprepared for the future if they are not taught how to use AI efficiently and ethically.

Students can only know the limitations of AI by using it. Ethan Mollick is not only allowing AI in his classroom, but he is mandating it. Mollick’s AI policy requires students to assume the AI is wrong, adding that students will be responsible for errors and omissions on behalf of the AI. Mollick also requires his students to include a paragraph at the end of their work disclosing the use of AI, as well as the prompts used.

Contrary to popular belief, turning in a research paper is not the height of academic evaluation. According to Kathy-Hersh Paseck, it is more important that students develop writing skills, like how to write a thesis and support it, rather than turning in a paper. Students are able to learn these necessary skills by interacting with AI.

Donnie Piercey’s fifth grade class plays a game of “find-the-bot,” which asks students to pick the AI summary out of a lineup with peers’ work.  The students said that the exercise helped them identify proper capitalization and punctuation, as well as how to correctly summarize information. This exercise also led to a discussion on writing voice, and why the AI text sounds “stilted.”

AI text generators are tools that should be embraced in the classroom. Like other technology that was new for its time, AI will not go away. The best thing educators can do for their students is to teach them efficient and ethical ways to use AI text generators while still honoring academic integrity and learning essential writing skills.