Antibiotic resistance exists a rapidly growing public health crisis, causing more than one million deaths annually worldwide and contributing to almost 5 million additional deaths. These infections are more tough and exorbitant to treat than typical infections and result in longer hospital stays, increasing costs for both hospitals and patients.
Treatment mostly comes down to doctors’ guesswork. Ara Darzi, a surgeon and director of the Institute for Global Health Innovation at Imperial College London, says artificial intelligence-based diagnostics are a better way to go.
“Right now, in 2026, we are at the first real turning point of the crisis,” Darzi told WIRED Health in London on April 16.
The overuse and misuse of antibiotics and the lack of development of novel drugs are driving the boost in microbial resistance. When bacteria are exposed to antibiotics that do not kill them immediately, they develop defense mechanisms to survive. Unnecessary prescriptions allow bacteria to develop resistance, rendering life-saving drugs ineffective. This means the list of treatment options for patients with grave infections is shrinking.
The problem will get worse. AND Report for 2024 in The Lancet predicted that drug-resistant infections could cause 40 million deaths by 2050.
Conventional diagnostics to determine an antibiotic-resistant infection usually take two to three days because it requires culturing the bacteria from the sample. However, for some infections, such as sepsis, this is time that patients do not have. Each hour of delay in treatment increases the risk of death by 4–9%. While waiting for test results, doctors must exploit their best judgment in selecting the antibiotics they exploit.
AI-based diagnostics can support with decision-making. “AI-based diagnostics achieve accuracy above 99 percent without additional laboratory infrastructure,” Darzi said.
He added that this type of rapid diagnostics is especially needed in rural and remote areas of the world. World Health Organization estimates that antibiotic resistance is highest in Southeast Asia and the Eastern Mediterranean, where in 2023 one in three recorded infections was resistant. In Africa, one in five infections was resistant.
Artificial intelligence can also support discover novel drugs for resistant infections and predict the spread of resistant bacteria. The British National Health Service is working with Google DeepMind to develop an artificial intelligence system to combat antibiotic resistance. In one demonstration, the system identified previously unknown resistance mechanisms in just 48 hourssolving a mystery that took researchers from Imperial College London a decade to understand.
Darzi said that when combined with the automated lab, it is now possible to run hundreds of parallel experiments around the clock. Deep learning models can now examine billions of molecular structures in a matter of days, while generative artificial intelligence is being used to design compounds that do not exist in nature.
However, vast pharmaceutical companies have abandoned development of antibiotics due to a broken economic model. Up-to-date antibiotics would need to be reserved to prevent resistance, but pharmaceutical companies profit from large-scale sales. Companies don’t have much incentive to stay in the game.
Darzi argued that novel payment models are needed to encourage the development of novel antibiotics. In 2024, the UK began piloting a Netflix-style payment model, in which the government pays a pharmaceutical company a fixed annual subscription fee for access to novel antibiotics, rather than per quantity prescribed. Sweden is also experimenting with a semi-disconnected model.
“The question that will determine the shape of medicine for the next 100 years is not whether we have the tools to respond. We have the tools,” he said. “The question is whether we have the character to take what we see seriously.”