Scientists use AI to search for new superbug-killing antibiotics

Natalie Clarkson
by Natalie Clarkson
22 June 2023

Artificial intelligence (AI) is having a profound impact on the world in many ways right now – one application being the discovery of a new antibiotic by scientists in Canada and the US that is effective against a menacing, drug-resistant bacteria.

Scientists at Massachusetts Institute of Technology (MIT) and Canada’s McMaster University have been working with non-profit Phare Bio to develop an AI model that can discover new forms of antibiotics. An Audacious Project grantee in 2020, Phare Bio has a goal of developing seven new antibiotics in just seven years – currently it can take as long as 15 years to find just one. 

The Audacious Project, housed at TED, was brought to life to identify and fund big ideas that tackle the world’s most urgent challenges. Audacious is a collaboration of some of the most respected names in the non-profit world, including the Skoll Foundation, the Bill & Melinda Gates Foundation, Scott Cook & Signe Ostby and more. Virgin Unite is proud to be one of the partners that helped create Audacious, which has allowed us to go on to support incredible organisations like Phare Bio. 

More than a million people a year are estimated to die from infections that don’t respond to treatment with antibiotics. But there has been a lack of new drugs for decades and bacteria are evolving to be resistant to the existing ones.  

“I think that [AI models], when appropriately trained and leveraged for various prediction tasks, can help increase the rate at which we discover new candidates for drug development,” Jon Stokes PhD, a professor at McMaster University, told Fierce Biotech Research

Phare Bio CEO, Akhila Kosaraju MD, added: “We believe that this is a tremendously powerful, efficient and accurate way to move drug development into a new era. It’s particularly well-suited for antibiotics, where you can quickly test what’s generated out of the AI in a petri dish and determine how efficacious those predictions are.” 

The researchers at MIT and McMaster University have recently focused on Acinetobacter baumannii, a bacteria that can infect wounds and cause pneumonia. It’s one of three superbugs identified as a ‘critical’ threat by the World Health Organization.  


Researchers trained the AI with information relating to thousands of drugs with known chemical structures and efficacy (having manually tested them on Acinetobacter baumannii to see which had an effect on it). They then ran the trained AI across 6,680 compounds with previously unknown antibacterial efficacy.  

In just an hour and a half, the AI had produced a shortlist of 240. These were then tested in the lab, which led to the discovery of nine potential antibiotics. One of them was the antibiotic abaucin, which in laboratory experiments was able to kill Acinetobacter baumannii samples from patients. 

It’s still a long road until patients will receive the first antibiotics discovered by AI – scientists expect it to take until at least 2030. But the work that Phare Bio and its partners at MIT and McMaster University are doing has the potential to change how drugs are developed in the future and save lives lost to antibiotic-resistant infections around the world. 

Visit Phare Bio to learn more about its work.