Wail conducted his research in collaboration with supervisor and university lecturer David Neal from the eHealth Living & Learning LabExternal link. They investigated how AI can support clinical decision-making and what questions students have about the use of AI-tools.

“My research focused on questions students have about the use of AI-tools in the Dermatology department. I chose dermatology because I am doing my first internship here. The insights I gain from my research I can immediately share with my colleagues,” Wail says with a smile. 

Reliable

Through a survey, Wail collected students' questions. “I categorized these questions into themes. Then I delved into the literature. Most of the questions were about reliability and accuracy. Can AI help detecting or rule out skin lesions? And how reliable and accurate is the diagnosis? The outcome of my literature search was positive. Using AI-tools in dermatology makes equivalent or better diagnoses than without AI tools.”

AI as a support-tool

Should we worry that AI will take over decision-making? Wail doesn't think so. “We should see AI as a tool that supports us, and not as something that replaces our human thinking ability. It is important not to trust AI blindly and to keep our own medical knowledge up to date. The use of AI is a process of continuous collaboration, where we need to keep learning from each other. Humans will always be needed to verify the outcomes of AI. If we use AI in the right way, it will improve decision-making.”

Confidence

Wail's research shows that most medical students view the arrival of AI with confidence. “Although AI will play an important role in medicine, it will not fundamentally change the field,” Wail believes. “Human care and AI are complementary. AI supports with analysis and diagnostics, while empathy and final decision-making remain in human hands. It is teamwork between humans and technology.”

Research phase

AI may be developing rapidly, but many clinical decision support tools based on AI are still in the research phase. “Before these tools can be used in practice, they must be tested on patients. It also needs to be examined whether the tools perform well across different patient groups and conditions. Only after these steps an AI-tool be can safely and effectively implemented in healthcare.”

Education

To prepare students well for using AI tools in practice, education on AI could already be introduced during the bachelor's of Medicine programme. “More evidence is needed to integrate AI into the curriculum, but an elective on AI or Medical informaticsExternal link would be a valuable addition,” says Wail.

If you have any questions about this research or would like to know more, please send an e-mail to the following address: wail.ajaid@student.uva.nl