To create a syllabus for the new class about artificial intelligence he’s teaching starting fall semester, LSU Vice President and Provost Roy Haggerty fittingly used ChatGPT in a move he believes saved hours.
The class, called HNRS 3035, is designed to teach students the ethics, real-world applications and development of large language models, a variety of AI akin to ChatGPT. The class was announced via email in late July.
Haggerty personally oversaw the course’s creation in response to the “world changing” after the widespread release of ChatGPT late last year.
“Tools existed prior to ChatGPT being released, but what ChatGPT did was democratize access to these kinds of models, meaning that all of a sudden, essentially the entire world had access.” Haggerty said.
He said he spent many weekends and evenings learning about ChatGPT under the pretext that large language models held tremendous potential to complement applications in higher education. Every part of day-to-day work can benefit from it, he said, faculty and students alike.
Haggerty explained this by providing a hypothetical he believes is a realistic possibility: He said the university could benefit from a large language model, which he dubbed MikeGPT for the sake of example, that could prowl LSU’s multitudes of publicly accessible documents to provide users with information that should be theoretically accessible, but are hard to locate or understand.
“It’s a very, very rapidly evolving field,” Haggerty said. “One of the best ways to help position LSU to take advantage of AI would be for us to have a real world class on AI here. Why I’m teaching it personally is because I’d like to develop the LSU capacity in artificial intelligence for practical use at the university.”
And Haggerty said he’s aware of the piling worries and criticisms that a large number of professors hold over these rapidly evolving technologies, many of whom saw their syllabuses’ plagiarism and academic dishonesty clauses disenfranchised overnight.
“The better way to look at these is ‘How can I teach the students to use these technologies responsibly,’” Haggerty said.
He said AI’s emergence in the classroom is similar to the introduction of the calculator when he was a child. It was true that because of calculators, doing long addition mentally was harder, he said, but it was more practical to use a calculator for the sake of furthering mental capacities in other ways.
Leveraging AI in the classroom is like the calculator, he said, adding it’s true that many large language models suffer unignorable flaws, like ChatGPT’s tendencies to fabricate references or make up information.
But by front-running the class, he said, LSU can trailblaze in an industry that’s here to stay.
“The reason I’m teaching this class is to try to seed the local economy with individuals who are knowledgeable about this topic. In the end of this we’ll have a dozen to 20 students, mostly graduating seniors, who understand how to deploy and develop large language models.”
James Ghawaly, a senior AI and machine learning research scientist at the Stevenson Disaster Management Institute, will be the class’s technically focused professor. Ghawaly said by the end of class students will have made working AI models usable by the university and local businesses.
Henry Hays, a disruptive innovation entrepreneur and AI consultant for business will be the class’s business focused professor. Hays said the response he received after contacting large local businesses seeking AI help through LSU’s class was overwhelming.
Ghawaly and Hays will do most of the class’s teaching, Haggerty said.
Ghawaly said the class will kick off with an introduction to machine learning, followed by how Python coding relates to large language models and a dive into concepts like neural networks and deep learning as they relate to AI.
Enrolled students are not required to have a background in coding or AI, Ghawaly said.
They will work in teams made up primarily of business and computer science students, Hays said, and the entire class is project based. There will be two projects directly related to LSU and there are two pending projects related to local businesses, he added.
Hays emphasized how important it was young professionals to know how these systems worked entering the workforce; he said it’s because they’ll be expected to know.
All three of the class’s instructors said that LSU was the only university they were aware of that offered a class of this caliber.
Going forward, all of them believe, to some degree, the class will continue as a mainstay course.
“This isn’t going away is the understatement of the year,” Hays said. “I suspect that the idea of this class, the collaboration, bringing in different resources and parts of the university to assist the students’ learning will only grow from here.”
