Fleming Initiative & AWS: Using Gen AI to Fight AMR

Medical researchers are deploying artificial intelligence to combat antimicrobial resistance (AMR), a health condition where microorganisms evolve to withstand treatments that previously controlled them. The Fleming Initiative, a collaboration between Imperial College London and Imperial College Healthcare NHS Trust, is building a global intelligence platform to address drug-resistant infections.

Amazon Web Services (AWS) is providing technical support worth up to several million pounds to the Fleming Initiative. The package includes cloud computing resources, generative AI technology and technical expertise for the platform development.

According to the Fleming Initiative, AMR could cause 39 million deaths between 2025 and 2050. The organisation aims to connect data sources, researchers, healthcare professionals and public health bodies to address this threat.

Fragmented systems limit progress

Progress in addressing AMR remains slow due to disconnected surveillance systems and isolated research programmes. Healthcare organisations face restricted access to integrated data from hospitals, laboratories and community health settings.

Professor the Lord Darzi of Denham, Executive Chair of the Fleming Initiative, spoke at the One Health Summit in Lyon during France's G7 presidency. He described the AMR environment as containing millions or billions of data signals across an interconnected ecosystem.

"Medicine and public health are increasingly driven by data," he says. "The opportunity now is not simply to gather more of it but to turn it into action; at the speed and scale this threat demands.

"By pairing world-leading scientific expertise with the most advanced technology available, we can build a new generation of intelligence for AMR: one that allows countries, researchers and health systems to anticipate threats rather than react to them. That is the ambition this moment requires."

Cloud platform unifies datasets

The Fleming Initiative is using generative AI and cloud services from AWS to bring together fragmented AMR datasets. This marks the first time these separate collections, including compound libraries and surveillance information, will be unified on one platform.

The system aims to reveal patterns that were previously hidden across institutional divisions and national borders. By connecting these networks, research findings will be accessible beyond single laboratory holdings.

Professor Alison Holmes, Director of the Fleming Initiative, says AMR cannot be solved by one institution, country or dataset alone. "The support from AWS could help us unlock new opportunities to bring together expertise, data and technology in ways that were not previously possible," she says.

"By supporting more connected and accessible data ecosystems, researchers and public health leaders could collaborate more effectively, move faster and generate new insights at the scale and pace that matches the urgency of the AMR crisis."

The platform creates infrastructure for a global AMR intelligence system that becomes more capable as additional institutions contribute data. It supports collaboration across borders for researchers, healthcare organisations, industry bodies and policymakers to accelerate development work.

The shared infrastructure also builds capacity for AMR surveillance, preparedness and response at local and regional levels.

AI compresses research timelines

Artificial intelligence is shifting healthcare from reactive responses to predictive approaches. Dr Rowland Illing, Chief Medical Officer and Director of Global Healthcare and Life Sciences at AWS, says the technology helps compress years of laboratory research into weeks.

"AWS' generative and agentic AI services allow siloed organisations to bring together compound libraries, clinical data and genomic sequences in a secure fashion while respecting regional data sovereignty concerns, to find new treatments and enhance surveillance for new resistance patterns," Rowland says.

He notes that AMR disproportionately affects regions with less technology access. "AWS' AI services can accelerate data collection from these regions by transforming paper-based content into usable data, reducing or eliminating manual data entry processes through AI and agents and improving access to computational capacity," Rowland says.

"AWS creates a single, secure, cloud-based research environment that unifies these disparate datasets for the first time, enabling real-time access, AI-powered analysis at scale and multi-institutional collaboration without the constraints of physical infrastructure or jurisdictional barriers, to improve health outcomes for all."

The company supports in silico drug discovery for researchers. Generative AI screens the Initiative's library of more than 100,000 compounds and generates molecular candidates that may work against drug-resistant pathogens.

AWS AI tools are also used for resistance pattern prediction. The system trains foundation models on global genomic and surveillance data to forecast where and when new resistance patterns may emerge, providing public health agencies with an early warning system.