Blab AI
University of Alberta Study Shows Generative AI Boosts Health-Professions Students Confidence
When a modest $25,000 grant from the Teaching and Learning Enhancement Fund turned into a year‑long experiment, the University of Alberta’s College of Health Sciences found that generative AI can lift student confidence.
The research, carried out across nine iterations in eight courses—from pharmacy to physiotherapy and kinesiology—surveyed roughly 1,900 students before and after participation. The study aimed to see how integrating AI tools into real‑world learning activities would affect health‑profession students’ self‑efficacy and attitudes.
At the heart of the experiment were “Gems,” personalized AI assistants built on Google Gemini. Designed to act as interprofessional learning coaches, Gems let nursing students, for example, ask about the role of medical laboratory scientists and how those professionals collaborate in patient care. The team cautioned that the information Gems supplied should be verified by faculty or practicing professionals, and that instructors must remain transparent about AI use and co‑create guidelines with their students.
The numbers speak for themselves. Overall AI self‑efficacy scores climbed an average of 15 points on a 0‑100 scale, while confidence in health‑profession‑specific, activity‑aligned uses rose by up to 20 points. The authors linked these gains to social‑cognitive learning theory, which holds that meaningful practice opportunities reinforce self‑efficacy—a key predictor of competence. Despite the higher confidence, students’ attitudes toward AI stayed steady, reflecting a “cautiously curious” stance that balances enthusiasm with concerns about over‑reliance and reliability.
Beyond the classroom, the team explored AI’s potential in evidence synthesis and clinical documentation. Rojin Adabdokht noted that generative AI can aggregate findings across multiple studies, highlight common themes, and flag disagreements—an asset for evidence‑based disciplines. In pharmacy courses, students used AI to answer drug‑information questions that are not strictly black or white, then cross‑checked the AI output against approved sources. Ehsan Misaghi observed that the AI coach tended to agree with students rather than challenge them, prompting learners to seek more critical feedback.
The project also extends into real‑world health‑care practice. Through the AI in Medical Systems Society—co‑founded by Misaghi—the team is evaluating the accuracy of AI Scribe, a tool that records clinical encounters. Led by Dr. Robert Hayward, the study compares AI‑generated notes with human‑written ones across several Alberta family‑medicine clinics, with the goal of determining whether AI Scribe can free physicians to focus more on patient care.
Looking ahead, Dr. Tim Konoval is developing a graduate certificate in AI and health, with applications opening for the September 2026 cohort. The program will teach health professionals how AI can contribute to the Quadruple Aim—improving patient experience, advancing population health, reducing costs, and enhancing provider well‑being—without delving into deep technical coding. The team also offers a free, asynchronous “Foundations of AI Literacy” module to build baseline AI skills for students and staff.
The University of Alberta’s initiative demonstrates how modest funding can produce high‑quality scholarship in teaching and learning. By combining rigorous research, practical classroom tools, and real‑world clinical studies, the team shows that generative AI can boost student confidence, support evidence‑based practice, and potentially alleviate clinician burnout—provided that critical AI literacy and human oversight remain central.
The study’s findings are already informing multiple upcoming publications, and the university plans to expand AI‑enhanced learning across additional health‑profession programs. The graduate certificate and AI Scribe evaluation are slated for launch later this year, while the Foundations of AI Literacy course continues to enroll participants. The research underscores the importance of transparent, critically informed AI use in health‑care education and practice—a lesson that will resonate as AI tools become more embedded in health‑care workflows.
The research, carried out across nine iterations in eight courses—from pharmacy to physiotherapy and kinesiology—surveyed roughly 1,900 students before and after participation. The study aimed to see how integrating AI tools into real‑world learning activities would affect health‑profession students’ self‑efficacy and attitudes.
At the heart of the experiment were “Gems,” personalized AI assistants built on Google Gemini. Designed to act as interprofessional learning coaches, Gems let nursing students, for example, ask about the role of medical laboratory scientists and how those professionals collaborate in patient care. The team cautioned that the information Gems supplied should be verified by faculty or practicing professionals, and that instructors must remain transparent about AI use and co‑create guidelines with their students.
The numbers speak for themselves. Overall AI self‑efficacy scores climbed an average of 15 points on a 0‑100 scale, while confidence in health‑profession‑specific, activity‑aligned uses rose by up to 20 points. The authors linked these gains to social‑cognitive learning theory, which holds that meaningful practice opportunities reinforce self‑efficacy—a key predictor of competence. Despite the higher confidence, students’ attitudes toward AI stayed steady, reflecting a “cautiously curious” stance that balances enthusiasm with concerns about over‑reliance and reliability.
Beyond the classroom, the team explored AI’s potential in evidence synthesis and clinical documentation. Rojin Adabdokht noted that generative AI can aggregate findings across multiple studies, highlight common themes, and flag disagreements—an asset for evidence‑based disciplines. In pharmacy courses, students used AI to answer drug‑information questions that are not strictly black or white, then cross‑checked the AI output against approved sources. Ehsan Misaghi observed that the AI coach tended to agree with students rather than challenge them, prompting learners to seek more critical feedback.
The project also extends into real‑world health‑care practice. Through the AI in Medical Systems Society—co‑founded by Misaghi—the team is evaluating the accuracy of AI Scribe, a tool that records clinical encounters. Led by Dr. Robert Hayward, the study compares AI‑generated notes with human‑written ones across several Alberta family‑medicine clinics, with the goal of determining whether AI Scribe can free physicians to focus more on patient care.
Looking ahead, Dr. Tim Konoval is developing a graduate certificate in AI and health, with applications opening for the September 2026 cohort. The program will teach health professionals how AI can contribute to the Quadruple Aim—improving patient experience, advancing population health, reducing costs, and enhancing provider well‑being—without delving into deep technical coding. The team also offers a free, asynchronous “Foundations of AI Literacy” module to build baseline AI skills for students and staff.
The University of Alberta’s initiative demonstrates how modest funding can produce high‑quality scholarship in teaching and learning. By combining rigorous research, practical classroom tools, and real‑world clinical studies, the team shows that generative AI can boost student confidence, support evidence‑based practice, and potentially alleviate clinician burnout—provided that critical AI literacy and human oversight remain central.
The study’s findings are already informing multiple upcoming publications, and the university plans to expand AI‑enhanced learning across additional health‑profession programs. The graduate certificate and AI Scribe evaluation are slated for launch later this year, while the Foundations of AI Literacy course continues to enroll participants. The research underscores the importance of transparent, critically informed AI use in health‑care education and practice—a lesson that will resonate as AI tools become more embedded in health‑care workflows.
