ChatGPT can generate, but can it create?

Human-generated responses to ChatGPT might discuss its ethics, debate the legal concerns, rewrite our academic integrity policies, and brainstorm ideas for implementing it in our classes. Amusingly, ChatGPT can generate these, too, given the right prompts. But what can ChatGPT not do? At present, ChatGPT cannot access video content or refrain from hallucinating references, though both of these shortcomings may not exist for much longer. ChatGPT cannot see, smell, hear, taste, or touch. It cannot conduct research, observe results, or perform demonstrations. It cannot interrupt; it waits. It cannot initiate; it responds. It cannot think; it computes based on algorithms. It cannot let its mind wander through the stages of creative process because it has no mind. In short, it simply generates.

The pedagogical implications are clear. If all our assignments need to be revamped in order to mitigate the use of ChatGPT to complete them, can it mean that our educational system has taught students merely to generate rather than operate within the wider context and to the full extent of creation? It is currently impossible to know with certainty what is happening inside a learner’s head. To assess, we ask students to generate an output that serves as proxy evidence for cognitive activity. Have we been confusing a generated output with creation all along?

What’s Next

The same technology sector that has led current AI innovations will continue to bring ideas to market. Imagine a cottage industry to guarantee written products are created by human intelligence. Local agents might offer “100% Human Intelligence” certification guarantees. Perhaps in the future, writers will be surety bonded as an additional layer of protection to reputations. The next edtech company could produce an assessment platform that enables easy oral video examinations. Textbook companies may already be planning to train a bot on a textbook to offer as a companion product. Imagine we, the adopters of those texts, could purchase such a textbook bot at the course level with a “personalization” option that allows us to further train it on our syllabus. I do not currently foresee a full in loco magister state of affairs. There is too much money to be made to replace us, the potential consumers of such tools.

What I hope to see is an explosion of studies in the scholarly literature of AIEd to inform sound classroom practice. I dream of a spontaneity movement in education that not only encourages but expects students to interrupt and initiate—and they learn to reawaken those natural urges. I yearn for a supratheory of learning. If we leave the generation tasks to ChatGPT, we can focus on implementing immediate and local changes to someday create futures like these.

Consider engaging in the scholarship of teaching and learning in your classroom. Partner with an instructional designer who will support and extend your efforts and who will think deeply with you outside of the boxes that can impede our visions. Together, we can explore alternative teaching models, strategies, and methods. Perhaps we’ll create a new educational paradigm.

Related:
5 ways educators can leverage ChatGPT