Members of our Community Editorial Board, a group of community residents who are engaged with and passionate about local issues, respond to the following question: School is back in session. And artificial intelligence is back with it. AI tools, such as ChatGPT, are already changing classroom learning both at CU and in K-12 schools throughout our community. Your take?
Over the past nine months, ChatGPT has surely been the most talked about large language model out there. For instance, you too may have learned that one of the exceedingly well-worn jokes of the Spring 2023 graduation circuit was, “ChatGPT wrote that part of my speech.”
AI tools now help us process large amounts of data in order to mimic human conversations, answer exam questions, write poetry, compose songs and more.
With the start of a new school year, there’s renewed attention to advantages and drawbacks relating to inevitable student use. Alongside positive takes about harnessing their power, there’s also concern about students using ChatGPT to cheat. I myself have been teaching in university classrooms for over twenty years so ChatGPT has been a topic of spirited discussion.
AI tools have forced many of us in education to take a fresh look at how we assess student learning. And that’s a good thing. Stale assignments stifle student engagement, no doubt. Educators must work to innovate with the innovations in classroom teaching. We can do this by asking nuanced and creative questions that continue to help students develop critical thinking skills, and inspire novel explorations. We can also continue to encourage them to share their unique and valuable perspectives rather than regurgitating chatbot outputs while we appraise their learning.
I do worry about the dangers of uncritical AI uses. With it could come a creeping inability to assess the veracity of claims and the accuracy of statements and facts if AI information yields are bad. Mis- and disinformation can more readily spread under those conditions. (That said, I’m more worried about adults than student-age populations in this regard.) While we humans still have decision-making influence, let’s work together to minimize shortcomings and maximize benefits for effective use of AI tools in schools.
Max Boykoff, mboykoff@gmail.com
I was thrilled when I got access to an IBM Selectric II, with its magical backspace/erase key, to type my master’s thesis on the esoteric topic of “eschatology and the ethics of genetic engineering.” I was even more delighted when the Mac Plus appeared in time to write my genetics doctoral dissertation. These new “information technologies” felt revolutionary, like robotic partners in the misery of dissertation writing. But I never dreamed for an instant that they would evolve to pre-empt my role in our relationship, i.e., synthesizing ideas and information in creative new ways to increase knowledge.
Today’s information technologies have advanced to the point where they can write theses and dissertations without recourse to a creative and quirky human mind. At least that is what the hype proclaims, and what the venture dollars are chasing. But is that “intelligence”?
Like most doctoral programs, Notre Dame requires candidates to teach one or more semesters under the direction of a professor. I got assigned a few semesters of genetics, mostly for pre-med students, who were less interested in scientific reasoning than in “what they had to know” for a final grade that would help get them into med school. When confronted with a test question that provided some data and asked them to make a hypothesis (and, the horror, to design an experiment to test that hypothesis), they reacted with outrage. They just wanted to memorize stuff and spit it all back out in some superficially connected way.
Current AI output feels more like pre-med memorization than creative human thinking, though it can look eerily similar in a classroom setting. It may get close enough to fool us, but will never quite be the same. So maybe instead of worrying about its takeover, we should work on changing teaching from “what” to think to “how.”
Fintan Steele, fsteele1@me.com
To quote my sister Laura, a Board of Education member in CT, “AI is here, and the job of the schools is to teach kids how to use it and in what circumstances it is appropriate to use.” As K–12 districts scramble to update digital citizenship rules and strategies, teachers are already using ChatGPT to plan lessons and adapt learning materials. Like all new technologies, AI is likely to both energize learning but also lead to laziness (and cheating). It’s going to be more challenging than ever for schools to ensure students are actively engaged with their learning and becoming critical thinkers. To stay ahead of the game, and prevent students (and perhaps themselves) from leaning on AI too heavily, teachers will need to be nimble and creative — with plenty of in-class, handwritten and interactive work. At home, we parents and guardians will need to pay closer attention, too.
Despite (or because of?) the fact that in my 20s and 30s, I made a living writing formulaic teen fiction, I get a bit depressed contemplating the infinite quantity of passably literate text being spewed out by ChatGPT, not to mention imagining a future where “school” is AI grading papers written by AI. In the age of texting and social media, writing already sometimes appears to be a dying art. But maybe weirdly the use of AI in schools will lead more kids to appreciate what real literature is all about. Teachers might use AI to better support the large numbers of students who struggle to care about reading or to read proficiently. With technology and literature being integrated in innovative ways, there could be a rebirth of interest in the humanities. As my niece observed, “AI can write novels and paint pictures, but the best art and novels are created by people.” Amen!
Diane Schwemm, parksidediane@gmail.com
Artificial Intelligence tools like ChatGPT have woven themselves into various aspects of our lives. ChatGPT excels at summarizing articles and structuring ideas. However, it can’t replace the human touch — injecting soul into writing, validating facts or critically analyzing a subject. As a scientist, I rely on it for research, grant proposals and even drafting and proofreading my opinion pieces. However, when tasked to pen an opinion piece on a localized topic in Boulder, the bot would lack depth and personal nuance. I invariably have to supply an initial draft with my thoughts, and then tweak the bot’s output to capture my voice and thoughts. I repeat this process three to four times. The AI structures my ideas, but it’s still me behind the final narrative. It’s a game-changer in terms of efficiency but not necessarily proficiency.
However, for students not yet adept in writing or critical thinking, AI tools might mislead. ChatGPT’s confident tone, even when occasionally erroneous, can create misconceptions. Drawing from college 101 computer programming classes, where students first learn core algorithms before using packaged tools, we can apply a similar approach to AI in learning. In the formative years, students should be anchored in the basics of writing, pacing and structure without AI intervention. Once they grasp these foundations, introduce AI as an efficiency tool.
Unfortunately, the final burden falls on the teachers, who are already overworked and underpaid. The era of generalized essay prompts is gone. For instance, instead of a broad essay on the “Incan Empire road networks,” tasks should prompt deeper analysis, like “Discuss the Incan Empire’s road systems and draw parallels with Colorado’s current road maintenance strategies. How does road maintenance policy affect your daily commute?” While students can solicit basic information from ChatGPT, formulating a comprehensive answer will necessitate a blend of their knowledge and AI aid. Intricate prompts force students to engage critically, even when aided by AI. By focusing first on proficiency, then on efficiency, we can ensure that AI serves as a useful educational supplement without compromising critical skills.
Hernán Villanueva, chvillanuevap@gmail.com
