While ChatGPT seems to have sprung up overnight to dominate current headlines, legislative discussions, and professional debates—ranging from doomsday predictions to practical celebrations related to improved efficiency with menial labor tasks—the introduction of this online Artificial Intelligence (AI) program can be traced back to 2018. And while it is just the latest in a seven decades-old history of AI and Generative Pre-trained Transformer (GPT) that can generate content like text, images, music, etc., in response to user prompts, there is no ignoring the flurry of both concerns and hopes for its use in a variety of professional fields.
As part of the recent rush to comment on ChatGPT, this article is not weighing in on the ongoing ethical debates, nor is it offering a comprehensive how-to guide, but rather simply sharing one professor’s initial use of the AI tool and how it successfully accomplished the learning objectives and expectations for one course assignment.
For me, ChatGPT is just another in a long line of technology-assisted tools that I accept may become a routine part of my professional/personal life. As a busy MotherScholar (Burrow, et al., 2020), I am already grateful for customer service chat bots that allow me to schedule kids’ doctor appointments while waiting in carpool pick-up lines, and for automated digital writing assistants that catch my grammar and spelling errors as I madly fire off a manuscript between committee meetings and fixing dinner. To me, ChatGPT holds the potential to help make my academic life easier so that I can get back to enjoying the personal moments that truly count. It is this technology-enabled ease and efficiency that I thought could benefit my teacher candidates (TCs) in my fine arts for elementary education course during their busy end-of-semester season. The assistance that ChatGPT generated for my TCs was the time-saving jumpstart most needed to invest more energy into elements of their capstone project that actually needed their creative human touch.
ChatGPT-generated history of generative AI and ChatGPT[1]
Generative AI “refers to a category of artificial intelligence techniques that involve the generation of new and original content, such as text, images, music, and more” (ChatGPT, personal communication, May 18, 2023). Early approaches in the field of generative AI are situated within 1950s-60s computer science and AI research and “aimed to mimic human expertise to generate outputs” (ChatGPT, 2023). In the 1970s-80s, methods of generative modeling approaches “were widely applied in various domains, including natural language processing and speech recognition” (ChatGPT, 2023).
Developed by OpenAI, “an artificial intelligence research laboratory,…the history of ChatGPT can be traced back to the original GPT (Generative Pre-trained Transformer) model, which was introduced” in June 2018 (ChatGPT, 2023). In June 2020, OpenAI launched ChatGPT as a practical application of the GPT-3 model, which demonstrated “potential for numerous applications, including chatbots, virtual assistants, content generation, and more…[while] ChatGPT focused specifically on providing human-like conversational experiences” (ChatGPT, 2023). “ChatGPT uses a large amount of written data…[consisting] primarily of articles, websites, and academic papers, to generate human-like responses to user prompts” and the common uses of ChatGPT currently mimic those of GPT-3 including “chatbots, generated responses to user questions, marketing ideas, and coding or programming debugging” (Center for Teaching & Learning, 2023). In academic settings, ChatGPT can be used to produce written essays, reports, and responses when supplied with the assignment prompt.
ChatGPT, itself, confirms that while “generative AI continues to advance rapidly with ongoing research and development,” it could not provide me “with specific and up-to-date headlines about ChatGPT [as its] training only goes up until September 2021” (ChatGPT, 2023). However, it did assure me that “ethical considerations, such as biases, fairness, and responsible deployment, remain important areas of concern as the technology evolves” (ChatGPT, 2023).
Current examples of ChatGPT
Given the long history of AI-powered tools, there are obviously those who are more informed and comfortable with being at the forefront of innovative uses. For instance, at the June 2023 ISTELive event, C.L.Ai.R.A.[2] will make history as the first AI mainstage presenter. On the other hand, there are those who approach AI with great caution, advocating for government regulation, including OpenAI CEO Sam Altman who shares concerns that generative AI could, without safety standards, “mislead people, spread falsehoods, violate copyright protections, and upend some jobs” (O’Brien, 2023). Additionally, there are individuals who are not sure how to correctly use the technology, like the Texas A&M Commerce professor who incorrectly used ChatGPT to analyze student papers for cheating resulting in inaccurate accusations and failing grades (Murney, 2023). I fall into the camp of those willing to try it in my courses if I offer guided user-instruction, discuss logical safeguards, manage expectations, and emphasize a general understanding of best practices.
Capstone assignment gets a little AI assistance
As an often early-adopter of technology apps and programs that present novel ways to replace, amplify, and/or transform (Hughes, Thomas, & Scharber, 2006) my elementary education TCs’ course work, I quickly jumped into “trying out” ChatGPT with a spring semester capstone assignment that had been completed in my fine arts for elementary education course for the past six semesters.
In my fine arts course, TCs spend most of the semester experiencing and creating art to develop appreciation of and confidence in creativity so that they will hopefully encourage flexibility, risk-taking, and artistic expression in their own K-6th classrooms. TCs already use technology for many of the course assignments; for example, Spotify to curate playlists to tell the story of a significant life experience, Code.org/dance to choreograph an animated dance, and VR-enabled websites to virtually visit museums.
The course’s culminating assignment is meant to give them one last opportunity to bring together all the fine arts knowledge/skills learned throughout the semester. In previous semesters, TCs have written a fairytale and then transformed it into a script, filmed it, and added sound effects. But when informal class discussions about ChatGPT revealed that most of them had not heard of the generative AI tool, I quickly pivoted and revised the assignment by adding instructions to use AI to write their fairytales. After all, I teach a writing methods course, but this course is not that; this is a fine arts course and students’ writing expertise (or often lack thereof) is not the focus of the assignment—the story is simply a vehicle by which to drive arts integration into one final project.
Professor and student benefits
Most students enjoyed the novel “magic” of the experience and expressed gratitude for how ChatGPT helped them jumpstart their assignment, finish the work more quickly, and ultimately, let them focus more on the arts integration component of the assignment. Like good future teachers, some expressed reservations about using this in their future K-6th classrooms due to safety concerns and not wanting to rob young students of creative writing practice. But, again, for my students, this was not a writing assignment.
For my purposes, the use of ChatGPT truncated the course assignment timeline and allowed me to focus on evaluating the arts integration instead of critiquing their story writing. Essentially, it was just one more technology-enabled tool that helped students complete a final project that was already full of technological elements: filming, .wav sound effects, digital animations, and now, AI-generated stories.
A realistic conclusion – for now…
While some may notice that I am advocating for the use of ChatGPT for the very type of assignment that many academics are most worried about—writing!—I again reiterate that this assignment was not in a writing methods course, but a fine arts course. So, by removing students’ worries over time spent trying to craft a well-written fairytale when they had received no writing instruction during the semester, ChatGPT helped clear the way for them to show off their imaginations with the transformation of the source material into unique, creative, arts-focused productions.
While academics, lawmakers, and even technology professionals have shared legitimate worries about how ChatGPT could promote plagiarism/academic dishonesty, misleading information, and other falsehoods, I found that college students were conservative and conscientious in their use of the AI tool where creativity and originality were not the prioritized outcome. Like all tools, we will need to consider best practices, but assignments like this one demonstrate that students can be guided to use AI to generate the base product while still maintaining expectations for students’ original creative expression. My brief and early experience with promoting ChatGPT for an instructor-articulated purpose in an assignment leads me to believe that just like calculators, Grammarly, and other technology-assistant tools, there is a place for ChatGPT in college classrooms—and professors should lead the way in guiding students, not hiding from it!
Dr. Lauren E. Burrow is a professor of education studies at Stephen F. Austin State University, a member of the University of Texas system. She primarily teaches elementary education pre-service teachers, using community collaborations like Service-Learning and field trips as her primary pedagogy. She is a MotherScholar to three adolescent children who inspire and help create innovation and imagination assignments and experiences for her courses.
References
Burrow, L. E., Cross, C. J., Olson Beal, H. K., & Smith, S. (2020). The Skits, Sketches, and
Stories of MotherScholars. The Qualitative Report, 25(12), 4245-4273. Retrieved from https://nsuworks.nova.edu/tqr/vol25/iss12/4
Hughes, J.E., Thomas, R., and Scharber. C. (2006) Assessing technology integration: The RAT – replacement, amplification, and transformation – framework [Paper]. Society for Information Technology and Teacher Education (SITE) 2006, Orlando, FL.
Murney, M. (2023, May 18). “Texas professor flunks students after ChatGPT claims it wrote their papers.” The Houston Chronicle. https://www.chron.com/news/houston-texas/article/texas-teacher-chatgpt-fail-18104772.php)
O’Brien, M. (2023, May 17). “ChatGPTCEO to congress: AI-regulating agency needed ASAP.” The Christian Science Monitor. https://www.csmonitor.com/USA/Politics/2023/0517/ChatGPT-CEO-to-Congress-AI-regulating-agency-needed-ASAP
Center for Teaching & Learning [CTL] (March 20, 2023). Exploring ChatGPT.
[1] Most of the historical overview section for generative AI and the more recent history of ChatGPT was generated by ChatGPT based on my prompt: “Provide a brief overview of the history of ChatGPT” and the follow-up prompt: “Add history of generative AI.” The AI-generated responses were then combined with comments from my university’s Center for Teaching & Learning website, re-arranged, and paraphrased by me to present this background section. The entire user-computer interaction and subsequent author editing took a total of three minutes to complete.
[2] C.L.Ai.R.A. is “an AI-powered avatar [designed by Create Lab Ventures] that showcases what’s possible with digital human design. Known as the world’s first bilingual, autonomous, Afro-Latina AI avatar, she’s focused on making the world a better place” (https://conference.iste.org/2023/).
