While artificial intelligence (AI) has been a relatively silent partner in higher education’s early warning systems, personalized learning platforms, and more for some time now, we might fairly say that ChatGPT is the big boom heard ‘round the university. The AI chatbot is taking many of us by surprise and startling more of us to attention, not in small measure by its charming, eager extroversion: it “talks” to us. What’s happening here? Is ChatGPT a threat? What happens next?
Diffusion of ChatGPT
ChatGPT has been quite the busybot, going to business school, law school, the office, Congress, and more. We are experiencing the unfolding of Rogers’ (1962) innovation diffusion in real time. Since OpenAI released ChatGPT to the public in research preview on November 30, 2022, we’ve been busy ourselves, curating links and disseminating our treasuries to each other. We’re also creating artifacts such as the Advancements in AI Timeline developed by the Center for eLearning Initiatives at Penn State Behrend. The twin goals of all of our awareness-building activities are to hasten the development of our individual and collective opinions about whether ChatGPT is aide or adversary and to decide our next steps accordingly.
The truth is that the AI is a flawed facsimile of human intelligence, but depending on the task before it, it can be a remarkably capable one. For that reason, we’ve been putting it to fledgling use. The danger lies in the risk of distilling our efforts into empty “best practices” rather than informed recommendations because we’re building the empirical evidence as we fly. Not everyone is aware that Artificial Intelligence in Education (AIEd) is a decades-old field of study. (The term artificial intelligence was coined in 1955 by Dartmouth Professor of Mathematics John McCarthy, and the International Journal of Artificial Intelligence in Education published its inaugural issue in 1989.) The historically limited capabilities of the field’s principal subject of study may have contributed to what could be characterized as its stunted growth, until most recently.
Now as the big boom continues to echo, we’ll begin to subject our theories to empirical examination, and as practical evidence mounts, we’ll then be better equipped to confirm whether we’ve decided correctly.
An “Objective” View
ChatGPT’s name refers to one type of neural network machine-learning model, but many different AI models “generate” new information as they respond to a prompt. The word generate is often found in lists of suggested verbs that many find useful for crafting learning objectives. Such lists are based on Bloom’s Taxonomy, a framework originally designed and later revised and updated to provide a common vocabulary for educational assessment. Interestingly, the highest level of Bloom’s Revised Taxonomy is to create. Under create on verb lists, generate typically appears as one of many possible ways to demonstrate creation. Both verbs have Latin roots; both impart a sense of bringing something into being that did not previously exist. However, generate conveys a cause-and-effect, even mechanistic, process while create is evocative of a sense of growth, of development, of invention, of imagination, of the extraordinary.
