Hi ,

I hope your day is going well.

It is hard to believe we’re already in mid-August. Some of the trees in my neighbourhood have started changing colour—scattered yellows replacing summer greens. It feels as though summer has passed by far too quickly.

Lately, I’ve been saying, only half-jokingly, that my expertise as a physician may soon become redundant, thanks to the remarkable AI tools now becoming available. As a techno-optimist, I am trying to figure out my role in this new era.

AI is beginning to transform healthcare in numerous ways—enabling faster and more accurate diagnoses, improved interpretation of medical images, smarter drug discovery, personalized treatment planning, and even the early detection of disease patterns from electronic medical records.

AI in drug discovery: Why new antibiotics matter

Over the past few decades, very few new antibiotics have been developed. Most of the newly approved ones are only minor variations of older drugs. This is a dangerous gap because bacteria are constantly evolving and developing resistance to existing antibiotics. Bacteria are becoming "superbugs".

Resistance develops when bacteria are repeatedly exposed to antibiotics—whether in people, animals, or the environment. These survivors adapt, share resistance genes with other bacteria, and spread. The overuse of antibiotics for relatively minor illnesses, along with the use of powerful “last-resort” antibiotics in livestock feedlots, accelerates this process.

"Superbugs" are a significant threat and are responsible for an estimated 1.27 million deaths annually.

Traditional antibiotic discovery has been slow, expensive, and limited. Screening thousands of physical compounds in the lab can take a decade or more to find a potential molecule, and many promising candidates fail in later testing.

How AI is changing the game

With deep learning, machine learning, and the computing power to analyze vast datasets, scientists can now:

- Virtually screen millions of molecules in days rather than years

- Predict their likely interactions with biological targets

- Explore many more chemicals than humans could ever test manually.

Recent breakthroughs in AI antibiotic discovery

Halicin and Abaucin are two recent examples of new antibiotics.

Halicin was discovered at MIT in just three days via deep learning. It is effective against highly resistant bacteria such as Mycobacterium tuberculosis and Clostridioides difficile.

Abaucin—Identified within a year, is an antibiotic targeting Acinetobacter baumannii, one of the most dangerous hospital-acquired pathogens worldwide.

While both of these are still in early research stages, they show the speed and novelty AI can bring to the process.

From lab to patients: the first AI-designed drug in Phase 2 trials

Beyond antibiotics, AI is also being used to create drugs entirely from scratch. One example is rentosertib, the first fully AI-generated drug to enter Phase 2 clinical trials for idiopathic pulmonary fibrosis (IPF)—a chronic, progressive lung disease where lung tissue becomes scarred and stiff, making it increasingly difficult to breathe. “Idiopathic” means the cause is unknown, and current treatments are limited, with no cure available.

A Phase 2 trial tests whether a drug is effective in patients with the condition, while refining the dosage and monitoring safety. If successful, it will proceed to larger Phase 3 trials before potentially being approved for use in a doctor's clinic.

What makes rentosertib remarkable is how quickly it moved from concept to human trials. Using AI, researchers designed, refined, and brought it to patient testing in under 30 months—significantly faster than the decade or more it typically takes for a new drug to reach this stage.

AI may be our best shot at staying ahead in the race against drug-resistant bacteria, and it is already showing promise in designing entirely new drugs faster than ever before.

Next week, I will share what AI tools I use regularly.

Best regards,

Shabnam

Was this email forwarded to you? Sign up to get future emails straight to your inbox

Dr. Shabnam Das Kar, MD

Functional Medicine Doctor

Tiny Habits Coach

Email: info@drkarmd.com

References:

1. Artificial Intelligence yields new antibiotic
2. Liu, Gary, et al. "Deep learning-guided discovery of an antibiotic targeting Acinetobacter baumannii." Nature Chemical Biology 19.11 (2023): 1342-1350.
3. Xu, Zuojun, et al. "A generative AI-discovered TNIK inhibitor for idiopathic pulmonary fibrosis: a randomized phase 2a trial." Nature Medicine (2025): 1-9.