You don’t have to search for long to find educators talking about ChatGPT these days, and the conversation is all over the place, from the excited and energized:
While we both joked that [ChatGPT] signaled the end of Western civilization, my colleague half-seriously suggested that the only way we might ever change the culture of the new American classroom was to begin to use this kind of technology as part of our instruction. As we both laughed at the suggestion, I suddenly wondered aloud, “Why not?”
And the perpetually panicked:
The claim that [teachers] will be perceived as “fossils” for their rejection of [ChatGPT] is as equally preposterous as claiming that people who derogate SUVs risk being perceived as anti-automobile. The rejection of a single popular web-based entity will hardly precipitate our profession’s fall from grace in the public’s mind.
To the self-assured and steady:
Many have heralded the death of education due to [ChatGPT]. In fact, [ChatGPT] just enhances standard educational techniques.
Of course, it’s understandable why everyone’s talking about ChatGPT, with its tremendous potential for both academic advancement and misuse. After all, education has never faced a challenge like this.
Right?
Well…not quite.
See, there’s a reason those perspectives above aren’t in quotation marks—they’re actually not about ChatGPT at all. 
In truth, the first perspective is about using social media in the classroom, published in 2010.
The second person was freaking out about Wikipedia—that quote is from a letter to the editor of American Libraries, published in 2008. (Leave it to a librarian to use a word like derogate.)
And the last one is about—wait for it—using multimedia CD-ROMs in a classroom, published in 1995. Now that would get you labeled a “fossil” today.
So if we’ve been here before, where do we go next? In the past, education has innovated.
Today there are entire college courses and even degree programs in social media management and marketing. Wikipedia’s community editing platform has grown more reputable; incorrect information and viral jokes are quickly remedied; and students have learned to be discerning about the information they read there and check the sources on the page. As for multimedia CD-ROMs…well, those have mostly gone away, but it’s hard to imagine a modern teacher refusing to use any multimedia at all.
With this in mind, Thierry Viéville, a senior researcher in Computational Neuroscience at INRIA (the French National Research Institute in Digital Science and Technology) and the co-director of the Smart EdTech master’s program at the Université Côte d’Azur, believes the whole conversation about ChatGPT is focused on the wrong question.
Rather than worrying about how ChatGPT might affect education, Viéville says we should be looking at the bigger picture. What are the challenges facing education—and even more foundationally, what does it mean to learn something new?
“For a machine, learning means adapting the parameters of a very general algorithm in order to provide an approximate optimal input/output transformation, given a huge set of examples, a reward after each action, or an a priori optimization criterion,” Viéville said.
That may sound complex to people who haven’t spent their career researching computational neuroscience, but Viéville says human learning is a deeper process, taking place at the phylogenetic, ontogenetic, episodic, semantic, and procedural levels.
In other words, if students haven’t developed the skills they need to understand AI-generated information at those deeper levels of human cognition—skills like critical thinking, analysis, synthesis, and linguistic fluency—then we need to rethink what success looks like.
“The value of writing has never only been in the final product, but in the process—in the cognitive work required to develop and complicate one’s own thinking over a process of revision,” said Dr. Owen Matson, a former English lecturer at Princeton University and professor at the Georgia Institute of Technology. “If students don’t engage in the writing process, they don’t develop those skills. It’s kind of like how we still need to take walks for our health, even if technology has provided other means of transportation.”
Viéville agreed, adding that the many facets of human learning provide intrinsic value to learning skills that can be replicated by an AI or a machine—like counting or driving. Learning these skills allows you to “develop your cognitive functions, understand how the machine performs in order to use the result appropriately, and have a fallback option if the AI or machine is no longer working,” Viéville says.
We’ve seen this already with ChatGPT providing incorrect information, citing imaginary sources, and being overwhelmed by web traffic. Just weeks after its launch, a computer science student at Princeton spent a few days over winter break developing a tool to detect AI-written language with 98% accuracy. Last month, ChatGPT’s developers announced they were moving the full tool behind a paywall and developing their own AI content detector.
Of course, AI will continue to advance, and there may come a day when ChatGPT or another tool is able to replicate human language more fluidly. However, education will live to fight another day—reports of its death have been greatly exaggerated for decades. In the meantime, Viéville said with a smile, ”I look forward to discovering the next digital gadget that will generate all this media chatter.”
————————————————————
Tara Patterson is the Learning Sciences Content Manager at Shmoop. Before joining the team full-time, she played the long game as a Test Prep/ELA contract writer for Shmoop for more than six years, and prior to that, she worked in public relations for one of the largest firms in the Southeast. She graduated summa cum laude from Belmont University with a degree in journalism and thousands of dollars in student debt, and she is currently pursuing her master’s in Smart EdTech at the Université Côte d’Azur.
————————————————————
Conversation Sparks
This is just the beginning of the discussion! Here are the big questions we're asking ourselves these days:
- What are the biggest challenges and needs in education today? Could AI help or hinder our ability to address those?
- Is there a difference between students using 1) a calculator on their math or science homework, 2) Wikipedia on their social studies homework, and 3) ChatGPT or other AI on their Language Arts homework? If so, explain the difference.
- What happens when a student doesn’t use ChatGPT or AI assistance, but a teacher suspects they have and disciplines them for it?
- What are the untapped skills students could develop in conjunction with AI (like the media literacy skills they’ve developed with Wikipedia)?
- What is the difference between being able to spit out a concept (what ChatGPT can do) and actually understanding that concept (what humans can do)?
- What does it mean to learn something new?
- How might the unconscious bias of AI developers and the information used to train the AI affect the responses it generates?
Feel free to use these questions to start a conversation with your students or colleagues!
————————————————————
Sources
Broz, William J. “Not Reading: The 800-Pound Mockingbird in the Classroom.” The English Journal, vol. 100, no. 5, 2011, pp. 15–20. JSTOR, http://www.jstor.org/stable/23047797. Accessed 1 Feb. 2023.
Gable, Craig. “Wikipedia ‘Call’ Lacks Merit.” American Libraries, vol. 39, no. 9, 2008, pp. 11–11. JSTOR, http://www.jstor.org/stable/25650082. Accessed 1 Feb. 2023.
Garza, Thomas. “Blogging and Tweeting and Chat, Oh My! Social Networks, Classroom Culture, and Foreign Language Instruction.” Russian Language Journal / Русский Язык, vol. 60, 2010, pp. 123–37. JSTOR, http://www.jstor.org/stable/43669178. Accessed 1 Feb. 2023
Matson, Owen. Personal interview. 2 Feb. 2023.
Ofgang, Eric. “What is GPTZero? The ChatGPT Detection Tool Explained by Its Creator.” Tech & Learning, 27 Jan. 2023. https://www.techlearning.com/news/what-is-gptzero-the-chatgpt-detection-tool-explained. Accessed 2 Feb. 2023.
Viéville, Thierry. Personal interview. 1 Feb. 2023.
Wallia, C. J. S. “Using Multimedia in College English: State-of-the-Art.” Technical Communication, vol. 42, no. 2, 1995, pp. 342–43. JSTOR, http://www.jstor.org/stable/43087911. Accessed 1 Feb. 2023.