Bacterial diseases have plagued humans for as long as we have existed, and they still kill millions of people. Antibiotics transformed modern medicine by giving us a powerful way to fight them. But today, that protection is fading.
Key messages:
- Antibiotic resistance is a major global health threat. It was associated with more than 4.7 million deaths worldwide in 2021, and more than 70% of disease-causing bacteria are now resistant to at least one antibiotic in use.
- Antibiotic misuse affects humans, animals and the environment. More than two thirds of antibiotics are used in food-producing animals, and many of these are the same as or closely related to those used in human medicine. Researchers use a “One Health” approach to trace resistant bacteria in both animals humans, and the environment.
- The development of new antibiotics is not keeping pace with resistance. The number of new antibiotic classes has dropped since 1987, partly because pharmaceutical companies have moved away from antibiotic research due to low financial incentives, high costs and scientific challenges.
- AI is helping to speed up antibiotic discovery, but it has limits. AI can predict the properties of existing compounds and design new molecules, but many AI designs are too hard to make in the lab, and biased or incomplete training data can lead to candidates that fail in laboratory or animal tests.
- Animal models remain essential to antibiotic development, even with AI. Candidates must be tested in the lab first, then in animal models that closely mimic human infections. This step-by-step validation is what ensures new antibiotics are safe and effective.
The rise of antibiotic resistance
Antimicrobial resistance, which happens when bacteria develop the ability to defeat the drugs designed to kill them, has become a major global health threat. It was associated with more than 4.7 million deaths worldwide in 2021. Over the last century, the excessive use of antibiotics has driven bacteria to develop resistance, and some drugs no longer work against certain infections. Today, more than 70% of disease-causing bacteria are resistant to at least one of the antibiotics in use.
If resistant bacteria keep spreading, infections will become harder and harder to control. We could be left defenceless against illnesses that we currently treat in just a few days. The consequences would be severe for both humans and the animals in our care, making it increasingly difficult to maintain good health and welfare. The World Health Organization (WHO) estimates that by 2035, treating resistant bacterial infections will cost more than $400 billion a year.
To prevent this, coordinated policies have been introduced worldwide. Their priorities are to reduce infections, keep existing drugs working, tackle the misuse of antimicrobials and develop alternatives.
Antibiotic misuse is a concern for all
Efforts have too often focused on antimicrobial misuse in human infections, leaving animals, plants and the environment out of the picture, although they contribute greatly to antibiotic resistance. Many bacteria found in animals can also cause disease in humans. Many antibiotics important for human health are also used in animals. And resistant microorganisms can be passed on from animals to humans.
In fact, more than two thirds of antibiotics worldwide are used in food-producing animals such as pigs, poultry and cattle, and many of these are the same as, or closely related to, those used in human medicine. About 70% of antibiotics eventually end up in the environment through agricultural, medical and animal waste.
For this reason, researchers have started tracing resistant bacteria across humans, the food chain and the environment. This “One Health” approach treats the health of people, animals and the environment as connected, and aims to find new ways to block the spread of resistance.
Where have all the antibiotics gone?
Unfortunately, the rise in antibiotic-resistant infections is outpacing the development of new antibiotics. The traditional cycle of antibiotic discovery and development, which historically kept pace with emerging resistance, is now breaking down. Since 1987, the number of new antibiotic classes has dropped. This is partly because pharmaceutical companies have shifted their focus away from antibiotic research, citing low financial incentives, high research costs and the scientific challenge of finding drug targets.
The medical and research communities recognise that, without new antimicrobials, infections caused by resistant bacteria will soon become a leading cause of illness and death. Fortunately, recent advances in artificial intelligence (AI) and computer-based drug discovery are beginning to reshape the field.
Harnessing AI for antibiotic discovery and design
Recent years have seen AI applied at multiple stages of drug development, from identifying targets to refining candidate compounds. It offers new ways to explore the vast range of possible molecules and biological targets more efficiently than traditional methods.
In antibiotic discovery, AI mainly follows two complementary strategies. The first is predicting the key properties of existing compounds. The second is designing entirely new molecules with desirable therapeutic features. Together, these approaches help streamline the antibiotic development pipeline, from first discovery to preclinical evaluation. They have sped up the identification of promising candidates in silico (computers), and sometimes in vitro (laboratory conditions) and even in vivo (animals).
However, AI also faces significant practical limitations. Many of the molecules it designs turn out to be impossible, or too hard, to make in the lab, which creates a gap between computer design and laboratory validation. In addition, AI models are limited by biases and gaps in the data they learn from. As a result, the candidates they generate often fail to show the antimicrobial activity predicted, whether in vitro or in animals.
AI cannot work without animal models
Integrating AI into antibiotic discovery workflows is undoubtedly a major shift in how antimicrobial drugs are developed, and it should ultimately reduce our reliance on animal models. Even so, animal models remain a crucial step in discovering and validating potential antimicrobial compounds.
Every candidate still needs to first be tested in vitro in the lab to establish how well it works against bacteria and whether it is safe. Successful candidates then move on to in vivo testing in animal models that closely mimic human infections. These include work to uncover how the drug works and which bacterial targets it attacks, as well as thorough resistance profiling to assess whether bacteria could develop resistance to it in the clinic. This step-by-step progression, from computer prediction to experimental in vivo validation, is essential for selecting safe and effective molecules.
AI has made discovering and selecting drug candidates more efficient, streamlined many steps in the process and reduced reliance on animal models. But animal testing remains a crucial step that safeguards the entire clinical antibiotic pipeline. Through such coordinated efforts, AI-powered antibiotic discovery can move from a promising technology to real-world solutions, reshaping how we fight drug-resistant infections in the 21st century.
FAQ
What is antibiotic resistance?
Antibiotic resistance (also called antimicrobial resistance) happens when bacteria develop the ability to defeat the drugs designed to kill them. It is a major global health threat, associated with more than 4.7 million deaths worldwide in 2021.
What causes antibiotic resistance?
The excessive use of antibiotics over the last century has driven bacteria to develop resistance. Misuse in humans, animals, plants and the environment all contribute.
How many bacteria are resistant to antibiotics?
More than 70% of disease-causing bacteria are resistant to at least one of the antibiotics currently in use.
Why does antibiotic use in animals matter for human health?
Many bacteria found in animals can also cause disease in humans, and many antibiotics important for human health are also used in animals. Resistant microorganisms can pass from animals to humans. About 70% of medically important antibiotics are used in food-producing animals such as pigs, poultry and cattle.
What is the “One Health” approach?
It is an approach that treats the health of people, animals and the environment as connected. Researchers use it to trace resistant bacteria across humans, the food chain and the environment, and to find new ways to block their spread.
Why are so few new antibiotics being developed?
The number of new antibiotic classes has dropped since 1987. This is partly because pharmaceutical companies have shifted away from antibiotic research, citing low financial incentives, high research costs and the scientific challenge of finding drug targets.
How is AI used to discover new antibiotics?
AI follows two complementary strategies. It predicts the key properties of existing compounds, and it designs entirely new molecules with desirable therapeutic features. Together, these approaches help streamline the development pipeline, from first discovery to preclinical evaluation.
What are the limitations of AI in antibiotic discovery?
Many AI-designed molecules are too hard, or impossible, to make in the lab. AI models are also limited by biases and gaps in their training data, so the candidates they generate can fail to show the predicted antimicrobial activity in laboratory or animal tests.
Will AI replace animal testing in antibiotic development?
Not for now. AI should ultimately reduce reliance on animal models, but animal testing remains a crucial step. AI selected candidates are tested in the lab first, and the sucessful ones then move to animal research. This safeguards the whole clinical antibiotic pipeline.
Why are animal models still needed if AI can design antibiotics?
Computer predictions must be experimentally validated. Animal studies assess how well the AI-selected drug works, how it acts on its bacterial targets and whether resistance could develop in the clinic. This is essential for selecting safe and effective molecules.
What could happen if antibiotics stop working?
Infections that are treated in a few days today could become increasingly difficult to control, affecting both humans and animals. The WHO estimates that by 2035, treating resistant bacterial infections will cost more than $400 billion a year.
Last edited: 18 November 2026 13:53