30 January 2023

ChatGPT: Educational friend or foe?

Kathy Hirsh-Pasek and Elias Blinkoff 

The invention of the telephone in 1876 was met with simultaneous amazement and trepidation. Critics wondered if phones would disrupt face-to-face communication in ways that made us either too active or lazy. When television entered our homes, we fretted about the potential harms of the box and screen time in every living room. Surely, this would create a society of couch potatoes who do not even notice the people sitting by their side and fail to engage in more important activities. The definition of “screen time” was later broadened to include the impacts of digital content and “social media” on children. Indeed, a recent article in The Atlantic by Professor John Haidt warns that the generation raised on social media could even imperil American capitalism and culture.

The latest challenge to the creative human intellect was introduced on November 30th, 2022 by OpenAI. ChatGPT is a conversational bot responsive to users’ questions in ways that allows it to search large databases and to create well-formed essays, legal briefs, poetry in the form of Shakespeare, computer code, or lyrics in the form of Rogers and Hammerstein, to name a few. As New York Times writer Kevin Roose commented, “ChatGPT is, quite simply, the best artificial intelligence chatbot ever released to the general public.”

Used in the right way, ChatGPT can be a friend to the classroom and an amazing tool for our students, not something to be feared.

As with the telephone, however, ChatGPT is primarily being met with amazement and trepidation. Some in education fear that students will never need to learn to write, as they can merely lean on ChatGPT. Writing for The Atlantic, English teacher Daniel Herman worried that ChatGPT spelled “The End of High School English.” In the same publication, Stephen Marche declared the college essay “dead.” Fortune Magazine quipped, “Is Chat GPT the end of trust? Will the college essay survive?” On January 3, 2023, the New York City Department of Education took the dramatic step of responding to these fears by blocking access to ChatGPT on all department devices and networks. A department spokesperson justified the decision due to “…concerns about negative impacts on student learning, and concerns regarding the safety and accuracy of content.” She further questioned the educational value of the technology, stating: “While the tool may be able to provide quick and easy answers to questions, it does not build critical-thinking and problem-solving skills, which are essential for academic and lifelong success.”

Educators, opinion writers, and researchers are engaged in a vibrant discussion about the implications of ChatGPT right now. The emerging consensus is that teachers and professors might be tricked. That is—ChatGPT would surely pass the Turing test. For example, Daniel Herman describes how the program drafted a reasonable college essay, a cover letter to serve as a manager at Starbucks, and even an academic paper comparing two texts. Microbiologist Alex Berezow further discovered that ChatGPT excelled at answering short-response questions from a college-level microbiology quiz. However, the essays produced by ChatGPT are still identifiable as bot-produced, rather than human-produced, due to a few fundamental flaws. The high school English paper that the program composed for Daniel Herman was superficial and lacked references. Other reports indicate that the program includes inaccurate information and fails to provide a compelling perspective, linking the writer and reader.

In our own test, the first author (Kathy) gave the bot a complicated essay question that she asks her Honors psychology students to answer. It did a respectable job. Yet—the bot produced no more than a B- or C+ essay. Why? To date, the bot cannot distinguish the “classic” article in a field that must be cited from any other article that reviews the same content. The bot also tends to keep referencing the same sources over and over again. These are issues that can be easily resolved in the next iteration.

More centrally, however, is that the bot is more of a synthesizer than a critical thinker. It would do well on a compare-and-contrast essay, but is less able to create a unique thesis and to defend that thesis.

As educators, we strive to make our students what John Bruer, former president of the McDonnell Foundation, dubbed knowledge transformers, rather than knowledge digesters. That means that memorization is less valued than critical thinking. In fact, one of the general problems with many educational systems today is that they value learning the facts more than being able to remember information over time, to generalize the learning to new situations and to creatively develop a new way of thinking about an issue. In a world in which all of the information since the beginning of time is said to double every 12 hours, memorization of facts quickly loses its currency.

HOW CAN CHATGPT CREATE KNOWLEDGE TRANSFORMERS?

The question before us is how we can productively use ChatGPT to help our students become knowledge transformers? A writer, a teacher, and an education professor all suggest an analogy from the calculator and math to ChatGPT and writing. In the same way that calculators became an important tool for students in math classes, ChatGPT has potential to become an important tool for writers who want to hone their critical thinking skills along with their communication skills. How might this happen? Educators are responding with valuable approaches. Adam Stevens, a high school history teacher in New York City who opposes his district’s decision to block ChatGPT, sees it as a valuable tool to promote—not limit—critical thinking. Students can evaluate the program’s initial response to a prompt, then consider how to improve it through revision. Other teachers quoted in a recent report on ChatGPT for Ed Week advocate for a similar approach and using the program to focus on the writing process. In higher education, we can openly let our students use ChatGPT for their class assignments, as well, and even use the bot in class to generate a first draft. Students can then learn how to move beyond the first draft to make their essays better. This is precisely the method that the first author will adopt at the start of her class after winter break.

DEEPER, MORE ENGAGED LEARNING

Our students already know how to use this new tool. They are likely more sophisticated than their teachers at framing the questions and getting solid answers from the bot, even though it was just released. What they need to learn is why—at least for the moment—ChatGPT would get a lower grade than they could get. It is exciting to see how quickly educators are responding to this new reality in the classroom and recognizing the instructional value of ChatGPT for deeper, more engaged learning.

As Adam Stevens remarks, ChatGPT is only a threat if our education system continues to “pursue rubric points and not knowledge.” It is critical for all educators to follow their colleague’s example. As we note in our recent book, “Making Schools Work,” the old education model in which teachers deliver information to later be condensed and repeated will not prepare our students for success in the classroom—or the jobs of tomorrow. We should allow that model to die a peaceful death. Used in the right way, ChatGPT can be a friend to the classroom and an amazing tool for our students, not something to be feared.

No comments: