An AI tutor developed at Iowa State University is delivering measurable improvements in student performance, with users scoring significantly higher on final exams and showing strong voluntary adoption rates. The tool, created over two years by assistant professor Karl Kerns and his graduate students, provides on-demand tutoring support that complements traditional classroom instruction in a large anatomy and physiology lab course.

What Makes This AI Tutor Different From Existing Learning Tools?

Unlike standard learning management systems such as Canvas, Kerns' AI tutor offers interactive feedback during guided practice sessions. When students answer a question incorrectly, the tutor provides real-time guidance rather than simply marking the response wrong. This capability mirrors the personalized support that instructors traditionally provide during office hours or lab sessions.

The tutor supports multiple learning approaches. Students can use it to review specific weekly content, generate flashcards for self-study, or engage in guided practice sessions to test their knowledge. The flexibility allows learners to access support whenever they need it, whether between classes, late at night, or on weekends.

"It's like having a Zoom session with a real-life human mentor," said Karl Kerns, assistant professor of animal science.

Karl Kerns, Assistant Professor of Animal Science, Iowa State University

How Did the AI Tutor Perform in the Classroom?

The tutor was introduced to students in ANS 2140L: Domestic Animal Anatomy and Physiology Lab during the spring 2025 semester as a soft launch, then fully deployed in the most recent semester. The course enrolls between 160 and 180 students each semester, making it an ideal setting to test scalability.

Results showed a clear correlation between tutor usage and academic performance. Students who used the AI tutor scored 4.6 percentage points higher on their final course grades compared to non-users. More significantly, students who engaged with the tutor four or more times finished with 9.1 percentage points higher on their final grade, roughly equivalent to a full letter grade difference.

Adoption was strong despite the voluntary nature of the tool. Approximately 40% of students in the course chose to use the tutor without being required to do so, suggesting genuine perceived value among the student population.

How to Implement AI Tutoring in Your Classroom

• Start with a Clear Learning Objective: Identify a specific course or topic where students struggle most, then design the AI tutor to address those gaps rather than attempting to replace all instruction.
• Build in Interactive Feedback Mechanisms: Ensure the tutor provides real-time guidance when students answer incorrectly, not just pass-fail scoring, to replicate the value of human mentorship.
• Offer Flexible Access Points: Design the tool to be available outside class hours so students can use it on their own schedule, reducing barriers to adoption.
• Gather Student Input Early: Conduct brief surveys to understand how often students are using the tool and whether it meets their learning needs, then iterate based on feedback.
• Position It as Supplementary, Not Replacement: Clearly communicate that the AI tutor enhances face-to-face instruction rather than replacing it, which helps address student concerns about AI in education.

Alex Else-Keller, a postdoctoral research associate in animal science who worked on the tutor's development, emphasized the importance of framing. "The goal for it is to be supplementary for people who need that face-to-face learning, not a replacement," she explained.

Else-Keller also noted that some students initially hesitated to use AI-based tools. However, once instructors provided clear guidance on how to use the tutor and students experienced its benefits firsthand, adoption increased. "Being able to give students instructions on how to use the tutor, then seeing them come back and tell me it was helpful was rewarding," she said.

What's Next for This AI Tutoring Model?

The Iowa State team is documenting their findings in a teaching manuscript that will detail the implementation process and outcomes. This research will help other institutions understand how to deploy similar tools effectively.

Kerns and his team received funding from the Center for Excellence in Learning and Teaching's annual Miller Fellowship program to develop the tutor. He hopes the tool can eventually be scaled across the entire university and made available through Iowa State Extension and Outreach, which provides educational resources to communities beyond campus.

The project reflects a broader institutional commitment to innovation in teaching. As Kerns noted, "We are always looking for ways to innovate in the classroom and to give students another useful tool. It's all about setting students up for success".

As Kerns