Faculty discuss recent developments in artificial intelligence

Photo Courtesy of Wikimedia Commons

If a journalist were to write an article on the impact of artificial intelligence (AI) on college, they might start it as follows:

“AI has transformed college essay writing and evaluation through advanced language models, instant feedback and personalized learning experiences.”

That would be a fine way to start the article, only a journalist didn’t write it. An artificial intelligence chatbot did.

ChatGPT, with the new model GPT-3, can generate written responses to prompts, including but not limited to essays, lists, computer code and language translation.

“Other companies have other chatbots. We just haven’t seen those be as accessible,” said Christopher Porter, director of the artificial intelligence program at Drake University. “I can’t access Google’s LAMDA and have it give me a response to a prompt I’d get in class, but with ChatGPT, I can.”

On Feb. 3, during the weekly faculty development meeting, Drake faculty held a panel to discuss the impact of chatbots on college essays. The meetings are often organized around current faculty areas of interest and concern, which lately have been centered around artificial intelligence.

“Literally all my faculty friends everywhere are talking about – what does this mean in my discipline, what does this mean for student learning, what does this mean for assessment?” said Deputy Provost Renee Cramer, who organized the event.

The panel was divided into three approaches to addressing AI: how to detect AI in writing, changing assessment of student progress and effectively using AI in the classroom. Over 40 faculty members registered for the event, which moved from Howard to the Cowles reading room to accommodate the crowd.

Cramer selected the steering committee for the panel based on their previous interest or expertise in the field. Members of the steering committee then staffed the panel. Drake faculty members Christopher Porter, Megan Brown, Clayton Mitchell, Judith Jones, Heidi Mannetter, Brett Sikkink and Renee Cramer, the panel’s originator, discussed the different approaches.

Unauthorized usage of AI at Drake University is currently categorized as a violation of academic integrity. As Drake currently leaves consequences for violations of academic integrity up to individual members of the faculty, that means that the consequences for unauthorized usage of AI depend on the professor.

“I don’t think we can talk about a ‘golden age’ of education that we need to return to,” Cramer said. “But there are skills that I know we want our students to gain capacity in that they might not gain with unreflecting use of these tools, just like they wouldn’t gain it with unreflecting use of any tools, including books.”

Professors that had utilized AI or detected it before discussed certain “tells” that the generated responses had with the crowd, before discussing projects that would prevent unauthorized AI use. Projects that discuss in-class material, are done in class or require exceptional nuance would be difficult, if not impossible, for an AI to generate.

Faculty members also discussed the root of why students might use AI, discussing the “Drake Busy” culture and intense grade pressure that many students today have.

Some professors have begun to explore ethically utilizing AI in their classes. Some approaches include training students to write effective prompts for chatbots or using AI-generated essays as starting points to criticize.

“One of the responses is, instead of saying ‘don’t use this tool ever,’ it’s ‘let’s figure out how to use this tool to advance what we already do,’” Porter said. “How can this help us with pre-writing, and now I can build upon this. How can I get ideas that I can put in my own voice?”

While some professors compared the introduction of AI to the introduction of Google or Wikipedia, others took a broader view of its effects.

“I think the scope of the problem is something maybe more akin to the introduction of the steam engine, as changing the nature of work,” Porter said. “You have manual labor, and now we’ve mechanized manual labor. So in the Industrial Revolution, that had cataclysmic consequences. Production went way up, but then a lot of people whose livelihoods depended on manual labor were not needed anymore.”

The panel did not address replacing human labor for AI, mainly focusing on AI’s impacts on colleges. 

There will be future faculty workshops about the impact of artificial intelligence, along with two panels on developments in ChatGPT from the angles of multiple disciplines led by Porter and Professor Martin Roth of Drake.