Scientists have developed an AI chatbot similar to ChatGPT to help governments develop effective policies to combat drug resistance.
Antimicrobial resistance (AMR), in which disease-causing bacteria and viruses no longer respond to the drugs designed to treat them, contributes to millions of deaths per year and results in healthcare costs as high as $412 billion per year, according to the World Health Organization. Health Organization (WHO).
In low- and middle-income countries, poor sanitation, limited access to quality medicines and inappropriate antibiotic use are fueling the rise in AMR, while serious conditions such as HIV, tuberculosis and malaria are becoming increasingly difficult to treat.
In 2015, WHO released a global action plan to tackle AMR under the One Health model, which recognizes the interconnectedness between people, animals, plants and their shared environment.
But “major gaps exist between ambitions and actions” when it comes to developing the necessary policies in low- to middle-income countries, according to a study published in the journal Environmental sciences and technology.
‘Smart friend’
The international team of researchers from the Chinese Academy of Sciences and the University of Durham, UK, have created an AI chatbot designed to bridge these gaps and help draw up national action plans.
The large language model tool, called the AMR-Policy GPT, contains information from AMR-related policy documents from 146 countries.
“AMR-Policy GPT is a conversational chatbot,” says David Graham, an environmental engineer at the University of Durham and lead co-author.
“It allows you to ask questions and get answers to the questions you ask.”
Unlike ChatGPT – which aggregates everything from the broader information universe to answer your questions – AMR-Policy GPT filters by quality, selecting technical information relevant to the topic, the researchers said.
“It’s like having a smart friend in the room,” Graham said SciDev.Net.
While the tool cannot formulate policy, it uses national policy plans, gray literature and other policy guidance from intergovernmental agencies to encourage lawmakers to consider different policy options, Graham says.
So if you are in a country for example in Sub-Saharan Africa and you have hardly any information about your own country, you can ask the bot and it will look around for information related to your question and your location.”
David Graham, environmental engineer at the University of Durham
“Ideally, the tool will provide decision makers with well-informed information from all disciplines, including livestock farming, crops, water quality and infectious diseases.”
Although designed with policy in mind, the tool can be used by anyone to ask questions about AMR. “And you don’t need to have advanced knowledge of AI,” Graham added.
As good as the data
Emmanuel Mukambo, a doctor and dementia researcher in Zambia, says AI is changing the way global health challenges are addressed by making information more accessible and easier to analyze.
“AI allows us to process large amounts of data quickly, discover patterns and gain insights we might otherwise miss,” he said SciDev.Net.
But he warned: “AI is only as good as the data it learns from.”
In the case of dementia, most of this comes from studies conducted in Western countries, Mukambo said, adding: “This makes it difficult to apply these findings to places like Africa, where research has not kept pace .
“AI tools have the potential to make a real difference in this part of the world, but only if we use them to amplify voices and stories that have been overlooked for too long,” he said.
Source:
Magazine reference:
Chen, C., et al. (2025). Using big language models to support antimicrobial resistance policy development: integrating the environment into health protection planning. Environmental sciences and technology. doi.org/10.1021/acs.est.4c07842.