My AIBot Did my Homework!

The future is here.

With the launch of ChatGPT in November 2022, students have access to a fine-tuned chatbot that can produce text-based answers to prompts ‒ essay prompts. The internet is abuzz with discussions regarding the future of education. ChatGPT’s writing is formulaic and simply mimics human tone; however, this development of AI technology is indicative of a transformation in modern academia. Are creativity and academic integrity on the line?

OpenAI, an AI research and deployment company based in San Francisco, developed ChatGPT. ChatGPT is a new, more advanced version of the original GPT-3 model. This program is unique because of its Generative Pre-training Transformer language generation model. Essentially, ChatGPT is trained on datasets of human conversations. The user may prompt ChatGPT with natural language, and because of its training, it will respond in a conversational way.

The bot acquires its knowledge from great volumes of information on the internet. It does not “know” anything but rather recognizes patterns on the internet to gather an accurate response. For example, you can ask ChatGPT classic Encyclopedia questions like “Who is Isaac Newton?” Or, you can give it a command such as “Explain the Civil Rights Movement in simple terms” or “Write a script for a science fiction novel.” The more specific you are in the prompt, the more precise the ChatGPT bot’s response will be.

Objectively, the ChatGPT AI Bot is revolutionary in terms of modern technology. The New York Times reported that 30 million people use ChatGPT daily. Due to its high user capacity, the bot is often unavailable. However, even with its rapid traction, to some, ChatGPT is an ill omen. 

In a recent article from The Atlantic, “The College Essay is Dead,” writer and former Shakespearean Literature professor Stephen Marche suggests that essay writing may become obsolete. Throughout the article, Marche contemplates the concerning clash between the humanities and technology in a world where STEM triumphs. He argues, “Neither the engineers building the linguistic tech nor the educators who will encounter the resulting language are prepared for the fallout.” Marche proposes that humanistic questions are “real questions with real consequences” that should only be answered and monitored by humans.

UK Educational Technology Professor Mike Sharples shares a different viewpoint, which Marche acknowledges in his article. In an article from The London School of Economics and Political Science blog, Sharples urges that if AI systems have a lasting impact on education, “that will come from educators and policymakers having to rethink how to assess students.” Sharples believes that teachers and students embracing new technology is inevitable, so teachers should consider creating assignments that machines cannot answer. ChatGPT may necessitate educators to think beyond essays. 

GT coordinator Ms. Bauer has hosted ongoing seminars regarding AI technology, specifically ChatGPT. Most recently, she gathered GT students studying AI and GT students from the Introduction to Education class to discuss the future of ChatGPT in schools. Srivatsa Kundurthy, Akshaj Mundada, Sakshini Ganesh, and Prisha Malik offered insightful information about the program, whereas Kyra Morena and myself asked questions with consideration to education. “The use of AI is extremely fascinating, but I also see the dangers ahead. I don’t think students will just use it for its informative purposes,” shared Kyra Morena in response to our discussion. 

Mr. Nugent was addressed in the conversation as an educator who has found a productive way to use the new AI technology. “It takes the tedium out of my portion of the job,” he explained. “It’s a great tool to establish the flow of my lessons.” To clarify, ChatGPT does not replace Nugent’s role as an educator. Rather, he uses it to create basic rubrics and flow charts. He has also found it particularly useful in finding sources for DBQs which would otherwise take a great amount of time on his part to assemble. “I do find it slightly unnerving,” he stated when discussing the advanced abilities of ChatGPT. “It helps teachers, but it could be dangerous for students.”

The good news? As explained by AssemblyAI, the creators use a technique called Reinforcement Learning from Human Feedback (RLHF) to minimize biased, harmful, or false outputs. Actual humans provide this feedback to ChatGPT. The bad news: ChatGPT still has a long way to go. The AI program is not yet in a state in which it can be controlled by task-specific prompts. ChatGPT is here to stay, and it is only growing as investors like Microsoft have fished out nearly $10 billion to ChatGPT developers. 

ChatGPT is simultaneously exciting and concerning. The future of AI in education is uncertain: not even ChatGPT itself has the answer.