A team of researchers from Inria Saclay, France, and the Indraprastha Institute of Information Technology Delhi (IIIT-Delhi), India, has introduced an artificial intelligence (AI)-based method designed to recommend alternative antibiotics for drug-resistant bacterial infections. The approach is aimed at aiding clinical decision-making by repurposing existing medications.
Antimicrobial resistance (AMR) continues to be a global public health concern. It occurs when bacteria no longer respond to antibiotics that were once effective. This makes routine infections, such as pneumonia, urinary tract infections, and minor wounds, more difficult to treat. According to recent studies, over 70 per cent of hospital-acquired infections in low- and middle-income countries are resistant to at least one common antibiotic.
The conventional drug development process for new antibiotics is slow and expensive, often requiring more than ten years and significant financial investment. As a result, healthcare professionals have turned to drug repositioning as a faster and more resource-efficient strategy.
To support this approach, the collaborative team led by Dr Emilie Chouzenoux (Inria Saclay) and Dr Angshul Majumdar (IIIT-Delhi) has developed a machine learning algorithm capable of suggesting alternate treatments for drug-resistant bacterial infections. The project team also includes research engineer Stuti Jain and graduate students Kriti Kumar and Sayantika Chatterjee.
The system applies a hybrid AI approach. Rather than depending solely on pre-defined rules or databases of antibiotic susceptibility, the model identifies patterns based on real-world clinical data. The researchers curated detailed antibiotic usage guidelines from leading Indian hospitals to reflect actual clinical treatment practices. These data were combined with molecular-level inputs, including bacterial genome sequences and chemical structures of antibiotics, to detect potential treatment options that are not widely utilised.
The AI system was tested through case studies involving multidrug-resistant strains such as Klebsiella pneumoniae, Neisseria gonorrhoeae, and Mycobacterium tuberculosis. These pathogens are known for causing hospital-acquired infections, sexually transmitted diseases, and tuberculosis, respectively. In each case, the AI tool proposed antibiotics with known or potential effectiveness. These suggestions were reviewed against existing resistance data and expert assessments.
“This is an excellent example of how AI and international collaboration can come together to solve real-world medical challenges,” said Dr Majumdar. “Our method makes it possible to use existing knowledge more effectively and opens the door to smarter, faster responses to AMR.”
The researchers note that the AI tool could be used in hospitals or public health settings to assist doctors and microbiologists in reducing treatment delays and improving antibiotic stewardship. The system is also expected to be relevant in contexts with limited diagnostic infrastructure.
The team sees this technology as part of future standard practices for infection management.
