Pitt faculty: AI reshapes grading and lesson prep
Math and English professors describe specific uses of large language models in their classrooms, ranging from generating exam problems to writing student feedback.
By Hugo Konate ·
A mathematics professor in Pittsburgh has trained an AI model to grade her midterms.
Anne Yust, a teaching associate professor of mathematics, told The Pitt News that she uses the tool to provide detailed feedback to approximately 150 calculus students. Yust stated that she traditionally spends about 40 hours grading one midterm. By using the software, she aims to reduce that time and spend more face time with students.
Yust said the technology now rarely makes mistakes on calculus problems, a major improvement over previous computational tools. She also used the AI to rewrite course policies to make them clearer and less overwhelming for students. Yust reported that students attended office hours more frequently to review their mistakes after receiving the AI-generated comments.
Other faculty members are adopting the technology for different tasks. Weifeng Xu, an assistant professor of neuroscience, uses AI to create lecture notes based on her slides. Xu fact-checks the generated text and tells students the notes are written by AI. Ross Harrison, an adjunct associate professor of political science, uses the tool sparingly to create charts and graphics for his lectures. Harrison compared the experience of using AI for research ideas to having a colleague sit next to him.
Maggie Jones, a teaching professor of English and contributing writer at The New York Times Magazine, uses AI for high-level brainstorming and finding primary source material. Jones said she found AI sources not always reliable and described the idea of using it for writing as "horrifying." She believes that journalism and personal essays must be the writer's own work.
Nick Canada, a graduate student in the English department, observed that faculty often use AI for organization and skimming large PDFs. Canada noted that a stigma around AI use in academia can lead some professors to hide their usage. He described the ethical line between appropriate usage and overreliance as difficult to define given the rapid development of the technology.
Yust hopes AI will continue to improve task efficiency so she can focus more on student interaction. Harrison suggested that the tool could be particularly useful for faculty members creating new courses. Canada said professors should use AI to help their ideas transfer better to students, but warned against it becoming a crutch for bad teaching.
