Will AI Improve Drug Development?
Developing a new drug is a long, expensive process, and many promising drugs fail before they ever reach patients. Artificial intelligence may help researchers find better drug candidates sooner and eliminate unsuccessful ones earlier.
Finding New Drug Targets
Before developing a drug, researchers first have to identify something in the body—a protein, gene, enzyme, or other biological process—that might be changed to treat a disease.
AI can analyze enormous amounts of genetic, molecular, and medical data looking for connections that researchers might otherwise miss. This could help identify new targets for diseases that currently have few effective treatments.
Designing New Drugs
Traditionally, scientists may test thousands or even millions of compounds looking for one that affects a particular target.
AI can help predict which molecules are most likely to work and even suggest entirely new molecules with desired characteristics.
Researchers still have to make and test these compounds in the laboratory, but AI may dramatically narrow the search.
Finding New Uses for Old Drugs
AI can also search existing medical and biological data for drugs that might work against diseases for which they were never originally intended.
This process, known as drug repurposing, is particularly interesting because much may already be known about an existing drug’s safety and side effects.
Improving Clinical Trials
Once a promising drug reaches human testing, AI may help researchers select appropriate patients, determine doses, monitor participants, analyze results, and design more efficient clinical trials.
AI may also help identify which patients are most likely to benefit from a particular treatment—and which might experience serious side effects.
This could be especially valuable for older adults, who are often underrepresented in clinical trials despite taking multiple medications and having several chronic conditions.
The possibility of developing treatments better suited to individual patients is one of AI's most interesting long-term promises.
AI-Designed Drugs Are Already Being Tested
This is no longer entirely theoretical. Drug companies are already using AI to identify targets and design experimental medicines, and some AI-assisted drugs have progressed into human clinical trials.
But an AI-designed drug still has to pass the same fundamental test as any other drug: Does it actually work in people, and is it safe?
An Example: An AI-Designed Drug for Lung Disease
One promising example is rentosertib, an experimental drug being developed for idiopathic pulmonary fibrosis, a disease that causes progressive scarring of the lungs.
Researchers used AI to identify a potential drug target and design a molecule intended to act on it.
In a small Phase 2a clinical trial published in 2025, researchers reported encouraging early findings, including a possible improvement in lung function.
The drug remains experimental, and larger studies are needed to establish whether it meaningfully improves patients' health and is safe over the long term.
This illustrates both the promise and the limitations of AI in drug development: AI may help researchers discover promising treatments, but only well-designed clinical trials can establish whether they actually work.
What Could Change?
AI probably will not eliminate the years of laboratory research and clinical testing required to develop a new medicine.
Its greatest contribution may be earlier in the process—helping researchers make better choices about which targets, molecules, and treatments are worth pursuing.
If AI can reduce some of the enormous amount of unsuccessful work involved in drug development, new treatments could eventually reach patients faster and at lower cost.
However, whether these advances will actually reduce the cost of developing drugs—or make new medicines more affordable to patients—remains uncertain.
Learn More
The FDA’s Center for Drug Evaluation and Research (CDER) explains how artificial intelligence is being used in developing medicines, including drug discovery, clinical trials, safety monitoring, and manufacturing.
A readable overview of how AI is being used to discover and develop new medicines, including finding promising compounds, predicting how drugs may behave, and improving clinical trials. It also discusses some of the limitations and risks.
A Phase 2a clinical trial of an AI-discovered and AI-designed drug. The results were encouraging but preliminary, and further testing is needed.