Two medical microbiology students of the Central University of Technology (CUT), in the Free State, recently presented a paper on the use of AI to identify pathogenic organisms at the University of Kent, in the UK.

As part of a diagnostic practical, they populated large language models with patient microscopy, culture and sensitivity data, supplemented with biochemical test results. They then critically evaluated AI-generated feedback.

Under the mentorship of CUT medical microbiology lecturer Professor Parks Makhoahle, third-year student Nthabiseng Tshandu and second-year student Chanelle Mcleod presented their paper.

Their assignment was focused on identifying bacterial pathogens exhibiting beta-haemolytic and weak alpha characteristics on blood agar, with specific gram morphological structures observed as long chains and diplococci/short chains respectively for different academic levels.

“Incorporating AI in learning and teaching has a potential to enhance student engagement, clearer understanding and better outcomes as much as it has its own challenges,” Tshandu said.

“Positively, AI can enable students to explore complex concepts in an engaging way, can tailor learning experiences to individuals’s needs and learning styles and also provides one-on-one support to students, offering real-time feedback and guidance.”

However, AI is not always accurate and reliable and biased data can lead to flawed decision-making, she cautioned.

The tool provided Tshandu with additional information by suggesting further molecular or biochemical testing to confirm the identity of the organism because it made tentative identification.

“If I had not used AI, I would not have known some concepts of Staphylococcus aureus which are similar to Enterococcus faecalis. It helped me in differentiating factors and same factors. It also enhanced my confidence because I had to go to other sources to compare my results with what I performed and I got to learn even more,” she said.

Further, AI can enhance learning and interpretation by providing tools for research, analysis and interpretation, McLeod said.

“It has helped me with my research, including the data analysis, interpretation of the microorganism and the antibiotic zone sizes I observed in my practical class. It also provided insights that improved my comprehension on this particular microorganism by providing a clear explanation to my results and findings.”

Further, AI helped confirmed her findings with more illustrations that deepened her understanding, elaborating on her explanation and interpretations, she said.

“It enhanced my ability to analyse microbial data by providing clear explanations for each piece of information. It also played a crucial role in simplifying microbiological processes. The AI tool also clarified laboratory techniques such as Gram staining, bacterial smear preparation, and heat fixation.

“Additionally, it provided insights into biochemical tests, antibiotic susceptibility testing, and antibiotic interpretation, enabling me to assess antibiotic effectiveness against bacteria,” she said.

“Our students showcased the integration of AI into medical microbiology diagnosis, emphasising its benefits while cautioning the attendees about the importance of information verification and human oversight in their final clinical decisions,” Makhoahle added.