Researchers at the University of Zurich (UZH) have used artificial intelligence (AI) to find bacteria that are resistant to antibiotics. Led by Professor Adrian Egli from UZH’s Institute of Medical Microbiology, the team is the first to explore how GPT-4, a powerful AI model by OpenAI, can be used to study antibiotic resistance.
The researchers used AI to help read results from a common lab test called the Kirby-Bauer disk diffusion test. This test shows which antibiotics can or cannot fight a bacterial infection. The scientists created a system called “EUCAST-GPT-expert” based on GPT-4, following strict EUCAST guidelines for finding out how bacteria resist antibiotics. They tested the system on hundreds of bacteria samples to see how well it could identify resistance to important antibiotics.
AI is Fast, But Humans Are Still More Accurate
“Antibiotic resistance is a growing problem worldwide, and we need faster and more reliable tools to detect it,” says Adrian Egli. His team’s research is the first step in using AI to help doctors find resistant bacteria more quickly. While the AI was good at finding some types of resistance, it wasn’t perfect. Sometimes, it mistakenly identified bacteria as resistant when they were not, which could delay treatment. Human experts were more accurate in checking for resistance, but the AI was quicker and could help speed up the testing process.
A Helpful Tool for Medical Staff
Even though the AI system has some limits, it shows great promise in healthcare. By giving a consistent way to read complex test results, AI could reduce the differences and mistakes that happen with manual readings, leading to better patient care.
Professor Egli points out that more testing and improvements are needed before this AI tool can be used in hospitals. “Our study is an important first step, but we’re not looking to replace human experts. Instead, AI can be a helpful tool for microbiologists,” he says.
Helping to Fight Antibiotic Resistance Worldwide
The study suggests that AI could support the global effort to tackle antibiotic resistance. With further development, AI tools could help labs around the world detect drug-resistant infections more quickly and accurately, helping to keep current antibiotics effective for longer.
Source: technologynetworks
