ROME, July 13 — The shift is happening faster than most people realise. Artificial intelligence is no longer just a tool for chatbots or image generators. The real transformation is unfolding in medicine, where AI is beginning to personalise treatments based on a patient’s unique genetic makeup, lifestyle, and environment.
This isn’t a future promise. It’s already in motion.
The trend underneath is straightforward: one-size-fits-all medicine is giving way to algorithms that can analyse vast amounts of biological data. In 2023, researchers at the Broad Institute used AI to predict drug responses based on the genomic profiles of cancer cells. That means a computer model can look at a tumour’s DNA and suggest which drug is most likely to work for that specific patient.
AI in the Lab and the Clinic
AI is also reshaping how clinical trials are designed. Instead of casting a wide net, algorithms are now identifying suitable patient populations and predicting trial outcomes before a single dose is administered. This could slash the time and cost of bringing new drugs to market.
The first AI-designed drug, developed by Insilico Medicine, entered Phase II trials in 2023. That’s a milestone.
A molecule conceived by a machine, not a human chemist, is now being tested in patients. Governments are paying attention. The National Institutes of Health launched the Bridge to Artificial Intelligence program, known as Bridge2AI, in 2022.
The investment: $130 million to create AI-ready datasets. The idea is that if you want AI to make sense of medical data, the data itself needs to be structured and clean.
That’s a massive infrastructure project, but it’s essential.
Regulation Catching Up
Regulators are moving, too. The European Union’s AI Act, first proposed in 2021, classifies medical AI as high-risk. That means any system used in diagnosis, treatment recommendation, or patient monitoring will need to pass conformity assessments before it can be deployed in Europe.
It’s a significant step — and a signal that lawmakers know this technology carries real stakes. In the United States, the FDA is expected to finalise guidance on AI device lifecycle management by 2025.
That covers how these systems are developed, validated, monitored, and updated over time. Because unlike a static drug, AI models can change as they learn. Regulators are still figuring out how to handle that.
None of this is simple. The challenges are real: data interoperability between different hospital systems, algorithmic fairness to ensure AI doesn’t reinforce existing biases, and integration into clinical workflows so doctors actually use these tools.
The technology works in a lab. Making it work at the bedside is a different problem. But the direction is clear.
AI is moving from research papers to real patients. The first AI-designed drug is in trials. The datasets are being built.
The regulators are writing the rules.
