This 2024 paper from the Journal of Clinical Microbiology explores the use of GPT-4-based AI agents to assist in identifying antimicrobial resistance (AMR) mechanisms such as ESBL, AmpC, and carbapenemase from disk diffusion tests.
Citation:
Giske et al., GPT-4-based AI agents—the new expert system for detection of antimicrobial resistance mechanisms?
Read full paper
https://journals.asm.org/doi/10.1128/jcm.00689-24
Highlights:
- Evaluated a customized “EUCAST-GPT-expert” using 225 Gram-negative isolates
- Compared against human microbiologists and non-customized GPT-4
- AI achieved high sensitivity (up to 100%), but lower specificity, especially for ESBL and AmpC
- Shows promise as a pre-screening support tool, though human judgment remains key
- Underscores the need for proper validation before clinical use in microbiology labs
What’s your take?
- Could AI assist your lab’s workflow or training?
- How do you see tools like this being adopted in practice?
- Would you prioritize sensitivity over specificity in early screening?