How did AI beat superbugs in two days?

This Article is sponsored by Phicus

Superbugs, those microorganisms that have eluded the effectiveness of antibiotics, have been a growing threat to global health for decades. Researchers around the world have spent years understanding their resistance mechanisms, a long and laborious process that has left modern medicine in check.

However, in an unexpected twist, artificial intelligence (AI) has achieved in 48 hours what science took more than a decade to discover: the secret behind its expansion. What seemed like a dead end is now shaping up to be the beginning of a new era in the fight against infectious diseases.

When Antibiotic Resistance Became a Global Crisis

By: Gabriel E. Levy B.

The problem of superbugs is not new.

In 1928, when Alexander Fleming discovered penicillin, humanity took a giant step forward in the fight against infections. But as early as 1945, Fleming himself warned that the abuse of antibiotics could generate resistant bacteria. His prediction was more than fulfilled.

In the decades that followed, microorganisms such as  methicillin-resistant Staphylococcus aureus (MRSA) and multidrug-resistant tuberculosis began to appear in hospitals around the world, challenging conventional treatments.

Today, antibiotic resistance has become one of the greatest threats to public health. According to the World Health Organization (WHO), in 2019 more than 1.2 million people died from antibiotic-resistant infections.

It is estimated that, if urgent action is not taken, this figure could exceed 10 million deaths per year by 2050. Medicine, with its traditional tools, has tried to contain this crisis, but the process of researching and developing new antibiotics is expensive and can take decades. In this context, the emergence of artificial intelligence in the field of microbiology represents unexpected hope.

Artificial intelligence enters the laboratory

Until recently, scientific research followed a classical method: observation, hypothesis, experimentation, and validation. This approach has been the foundation of modern science, but it has also been a slow process. However, the advent of artificial intelligence has changed the rules of the game.

Professor José R. Penadés, from Imperial College London, and his team had been trying for years to discover how some bacteria acquired resistance to antibiotics and managed to spread between species.

Their hypothesis, developed after years of study, suggested that these microorganisms could form tails from different viruses, allowing them to move from one host to another. It was an unpublished finding, not published anywhere.

When Penadés tested this hypothesis with “co-scientific”, an AI tool developed by Google, the surprise was huge. In just two days, the AI not only confirmed the team’s hypothesis, but also proposed four additional theories, all of which made biological sense.

Most impressively, one of them had not even been considered by researchers, which opened up new lines of study.

The end of traditional research?

The speed with which AI solved a problem that took scientists years raises an uncomfortable question: Are we facing the end of the scientific method as we know it?

Artificial intelligence has proven to be capable of analyzing large volumes of data, identifying patterns invisible to the human eye, and formulating hypotheses in a matter of hours. This could drastically reduce the time needed to make discoveries in biology, medicine, and many other disciplines.

However, the researchers insist that AI will not replace scientists, but will function as a complementary tool.

“It’s like having an extremely intelligent colleague who works tirelessly,” Penadés explained in an interview with the BBC.

Experimental validation remains essential, as AI can only generate hypotheses; Confirmation of these still requires real-world testing.

The ethical question also comes into play.

If artificial intelligences can solve scientific problems so quickly, who will guarantee that these advances will be used for the common good? Private companies such as Google are leading the development of these tools, raising questions about access to technology and its impact on scientific equity. Will developing countries be able to benefit from these innovations or will they be left behind in the technology race?

Cases that anticipated this revolution

The use of artificial intelligence in medical research is not an isolated phenomenon.

In 2020, Stanford University’s computational biology lab used AI to identify a new antibiotic, called Halicina, capable of killing resistant bacteria. What’s remarkable about this case is that the compound was discovered in just 48 hours by analyzing databases of molecular structures, a process that would normally have taken years.

Another notable case occurred in 2021, when the company DeepMind, a subsidiary of Google, presented AlphaFold, an AI capable of predicting the structure of proteins with unprecedented accuracy.

This breakthrough solved a problem that scientists had been trying to figure out for more than 50 years and has profound implications for the development of drugs and gene therapies.

These examples suggest that we are at the beginning of a transformation in science. While artificial intelligence will not replace critical thinking or human creativity, it could eliminate one of the main barriers to research: time. With the ability to generate hypotheses in a matter of hours and analyze data in unthinkable volumes, AI could accelerate medical discoveries, change the paradigm of drug development, and potentially save millions of lives.

In conclusion, the case of superbugs and artificial intelligence shows that science is entering a new phase.

What used to take decades can now be solved in a matter of days, thanks to AI tools capable of analyzing information at a speed unattainable by humans.

However, this does not mean the end of traditional research, but rather the beginning of an unprecedented collaboration between human and artificial intelligence.

As these technologies advance, the question is not whether they will change science, but how to ensure that their benefits reach all of humanity.