AI & Tutoring at Duke

The photos used in this publication are AI-generated.

Duke’s group of Peer Educators recently attended an AI training session led by the Academic Resource Center. The session provided valuable insights into how students can effectively incorporate AI tools into academic support roles. Duke’s Academic Resource Center employs over 80 student Peer Educators. At the time of this publication, Peer Educators have tutored over 700 Duke students this year in a variety of courses and disciplines.

Peer Educators are trained early each fall semester and participate in continuing education training, like this AI session, throughout the year. Training leader and Peer Educator Manager, Cheryl Beierschmitt, emphasized AI's role as a supplemental resource rather than a replacement for traditional learning methods.

“We can’t ignore AI and its impact on college students today,” noted Beierschmitt, “Giving our Peer Educators the best practices and tips for navigating AI allows them to responsibly incorporate AI into their tutoring work.”

Image

During the workshop, Peer Educators engaged in an AI prompt challenge where they collaborated to create the most effective prompts for generating study materials, explaining complex concepts, and designing interactive activities for students. They explored the potential of these tools to enhance tutoring efficiency and the importance of maintaining academic rigor by verifying AI-generated content against trusted sources. A significant portion of the training focused on addressing AI's limitations, including its tendency to produce inaccurate information.

“Generative AI is known for its potential of ‘hallucinating’ some information, from actual facts to the sources of information,” shared training attendee Lukas Bleichner, a Peer Educator preparing students for the MCAT, “It’s important to ensure students do not rely on AI passively without understanding the material.”

Peer Educators learned techniques to mitigate these risks, such as using AI to generate practice problems with intentional errors and to simulate challenging student questions or mistakes. The session also covered ethical considerations regarding data privacy and copyright, providing guidelines for responsible AI use in educational settings.

“I plan to apply AI tools like ChatGPT or Co-Pilot in my role as a MCAT Peer Educator to generate study guides and summaries of key topics tested on the exam, provide explanations for how to approach certain problems from different perspectives, and importantly, to help plan out and organize study schedules over a multi-month period,” Bleichner explained.

This practical approach demonstrates how AI can serve as a powerful organizational and explanatory tool when used strategically. The advice Bleichner gives to fellow Peer Educators serves as a message to everyone implementing AI in their work, “Use AI as a tool to clarify and reinforce learning, not as a shortcut.” Bleichner emphasized, “Encourage students to think critically and engage with the material rather than overly relying on AI to ‘think’ for them.”

Peer Educators are managed by the Academic Resource Center and are recruited each spring for the following academic year. For more information or to apply to be a Peer Education visit: https://arc.duke.edu/peer-education/.