0
0
New Delhi, May 23, 2025 — In a significant leap forward in the fight against drug-resistant infections, a team of researchers from the Indraprastha Institute of Information Technology-Delhi (IIIT-Delhi) and France’s Inria Saclay has unveiled an artificial intelligence (AI)-powered system designed to recommend effective combinations of existing antibiotics to battle so-called “superbugs.”
Superbugs are bacteria and fungi that have evolved resistance to multiple antibiotics, making infections increasingly difficult—and sometimes impossible—to treat. This growing threat, known as antimicrobial resistance (AMR), has been identified as one of the most urgent global health challenges of the 21st century.
A Collaborative International Effort
The AI project is part of a broader India-France research collaboration, involving Deep Light (Delhi) and CentraleSupelec, a prestigious French engineering school. The initiative is jointly led by Professor Angshul Majumdar of IIIT-Delhi and Dr. Emilie Chouzenoux of Inria Saclay, with key contributions from engineer Stuti Jain and graduate researchers Kriti Kumar and Sayantika Chatterjee.
“This is an excellent example of how AI and international collaboration can come together to solve real-world medical challenges,” said Prof. Majumdar. “Our method makes it possible to use existing knowledge more effectively and opens the door to smarter, faster responses to antimicrobial resistance.”
How the AI System Works
Unlike traditional rule-based models, the new AI system analyzes real-world clinical decisions from leading Indian hospitals, integrates bacterial genome data, and considers the chemical structure of antibiotics. This enables the tool to suggest optimal antibiotic combinations tailored to specific bacterial strains.
The system has already been successfully tested on multi-drug resistant strains such as Klebsiella pneumoniae, Neisseria gonorrhoeae, and Mycobacterium tuberculosis—all of which are known to cause severe, hard-to-treat infections.
Prof. Majumdar highlighted a particularly harrowing case from AIIMS, Kalyani, where a young patient’s artificial hip joint became infected and did not respond to even the most potent antibiotics. “It was a simple muscle infection, but it became untreatable. The patient from a poor family was left bedridden and without options. This is the reality of AMR—it can devastate lives even when the infection seems ordinary,” he explained.
Customizable and Scalable Solution
The AI model offers combination therapy suggestions, mirroring the approach doctors use to treat advanced infections. “Instead of recommending one antibiotic, our system can propose a cocktail of drugs based on the genome sequence of the bacteria,” said Prof. Majumdar. Doctors can request multiple options for a specific strain, and the AI will suggest viable treatments.
While the current focus is on bacterial infections, the system can be retrained to address viral infections or even lifestyle diseases like hypertension, where drug resistance is also emerging.
“Our core motivation was to address infectious diseases—a problem that plagues countries in Southeast Asia, Africa, and Latin America far more than it does the West,” Majumdar noted. “There’s less funding for this kind of work, but the impact is massive.”
Looking Ahead
The research team hopes to see their AI model integrated into hospital systems and public health frameworks, especially in regions with limited access to advanced diagnostics. As AMR continues to threaten global health, such AI-powered solutions could become vital allies for doctors, helping to preserve the effectiveness of antibiotics for future generations.
Disclaimer:
This article is based on information available as of May 23, 2025, and is intended for informational purposes only. The AI tool described is in the research and testing phase and is not yet widely available for clinical use. Readers should consult healthcare professionals for medical advice and treatment. The effectiveness and safety of the AI tool will depend on further validation and regulatory approval.
Citations:
