AI is giving longevity research a new way forward: better tools to identify potential interventions, and better ways to investigate their effects in people. Two recent papers from our team and academic collaborators illustrate that progression. In Cell , we introduced open tools to evaluate AI models and help researchers identify potential drug targets in aging biology.
In Nature Biotechnology , we explored how blood-based aging measurements could be incorporated into a clinical trial of an AI-developed drug. We cannot yet say the drug will extend anyone’s life, but these studies point toward a promising future for drug development: using AI to identify potential therapies and studying their effects on aging biology alongside the diseases they are designed to treat.
For business leaders, the opportunity is to help build that connection. AI can expand what scientists discover and measure; rigorous experiments and clinical trials must establish which advances improve health. That is how today’s research tools could become the foundation for tomorrow’s medicines.
I believe AI may eventually help double human lifespan. But that is an ambition, not a forecast we can validate today. The immediate opportunity is to build better medicines and the infrastructure to test them.
DRUG DISCOVERY RUNS ON TWO CLOCKS
The two clocks drug discovery relies on are computational and biological. The computational clock moves quickly. AI can analyze biological data, identify potential drug targets, generate chemical structures, and help scientists decide which experiments to run. It expands the range of possibilities researchers can explore.
The biological clock moves differently. A molecule must be synthesized and tested. Researchers need to establish whether it reaches the intended tissue, acts as expected, and causes harmful effects. Human studies require patient recruitment, dose selection, and sufficient follow-up to assess safety and benefit.
Our experience with rentosertib illustrates the distinction. The drug was developed for idiopathic pulmonary fibrosis, a disease that progressively scars the lungs. AI helped identify a biological target and design a molecule against it. We reached a nominated preclinical candidate in 18 months. That was a discovery milestone, not a finished medicine.
Laboratory studies and clinical trials followed. A Phase IIa study reported an encouraging lung-function signal, and the program has now advanced to Phase III testing . Each stage answers questions that the earlier stage cannot.
For executives and investors, this means measuring progress against the right milestone. A faster prediction, a promising experiment, and a successful clinical trial are different achievements. Putting them on one timeline can create unrealistic expectations about when patients will benefit.
A YOUNGER NUMBER IS NOT A LONGER LIFE
The aforementioned Nature Biotechnology analysis applied six proteomic aging clocks to blood samples from a subset of rentosertib trial participants. These models estimate biological age from patterns of circulating proteins. They detected treatment-associated shifts toward younger predicted profiles.
The result deserves further investigation. It does not mean patients literally became years younger. The analysis was exploratory, involved 42 people with a serious lung disease, and covered a short observation period. A change in proteins may reflect several processes, including changes related to disease.
The distinction matters beyond our company. As more businesses offer biological-age measurements, customers and investors need to ask what a changing score predicts. Does it correspond to better physical function, fewer illnesses, or longer survival? Has that relationship been demonstrated for the intervention being tested?
A biomarker can be a useful research tool before it becomes an accepted substitute for a clinical outcome. Treating those two roles as interchangeable risks turning an informative signal into a promise the evidence cannot support.
AI CAN IMPROVE TRIALS
There is considerable room to make clinical development more efficient. Better protocol design, more effective recruitment, and faster data review could reduce avoidable delays. AI may help researchers select useful measurements and identify patients most likely to benefit.
But efficiency cannot answer every biological question. A rare adverse effect may emerge only after enough people receive a medicine. A durable benefit requires follow-up. Whether a preventive treatment helps a healthy person decades later is an especially demanding question.
The practical response is to separate delays that technology can remove from observations science still needs. Invest in better trial operations, but also in the long-term studies and shared datasets that make their results meaningful.
Academic researchers are central to that effort. Their work on mechanisms, measurement, and experimental validation gives drug developers hypotheses worth testing. Partnerships should support that foundation as well as the visible commercial milestones.
BUILD FOR CUMULATIVE PROGRESS
For companies, the next step is to define success before making a longevity claim. Specify the population, the outcome, and the evidence needed to establish it. Publish methods and limitations. Distinguish exploratory findings from results a larger study has confirmed.
For investors, ask whether a company can repeatedly turn its predictions into credible experiments and clinical evidence. The number of molecules a model generates matters less than the quality of the decisions that follow.
I remain optimistic about radical life extension. I expect progress to come through successive generations of medicines, diagnostics, and other technologies that prevent disease and restore function. AI can help us develop and combine those advances.
Our own program offers an encouraging clinical signal and a new research question about aging. The next task is to test both more rigorously. The ambition can be extraordinary while the milestones remain concrete: better experiments, stronger evidence, and meaningful benefits for patients.
Alex Zhavoronkov, PhD, is founder and CEO of Insilico Medicine.
