Controlling aging
No treatment has been shown to slow or reverse human aging, the biggest risk factor for most diseases.
Open in the interactive tree →Aging is the gradual loss of repair and resilience that drives cancer, dementia, heart disease and frailty. Drugs and genetic changes extend the lives of worms, flies and mice, but human proof is missing, and regulators do not treat aging itself as a disease to be treated.
As of October 2026
The longest verified human lifespan is still Jeanne Calment's 122 years and 164 days (1997), and the FDA stated in 2023 that no medication has been proven to slow or reverse aging. A first human test of partial epigenetic reprogramming began in 2026: Life Biosciences received FDA clearance on 28 January 2026 for its ER-100 trial in optic neuropathies, dosed the first patient on 9 June 2026 and presented first-in-human data on 1 October 2026 (company statements). The 101 million US dollar XPRIZE Healthspan competition named 20 finalist teams in August 2026, which must restore muscle, cognitive and immune function by at least 10 years.
What is missing
- Validated biomarkers of biological age that respond to treatment and that regulators accept as trial endpoints
- A clear causal map of aging: which of the hallmarks are drivers and which are consequences
- Long and large human trials, since human lifespan makes tests slow and expensive
- Safe ways to reprogram or clear old cells without cancer risk
- A regulatory pathway that accepts aging itself as an indication
Becomes possible once solved
- Healthy lifespan beyond 100
- Preventing many age-related diseases at once
- Cell rejuvenation therapies
Open steps
- Age clocks regulators accept Medium AI leverageBiological-age measures that respond to treatment and predict real health outcomes, so a trial can use them as an endpoint.
- Which aging hallmarks drive aging Medium AI leverageSeparate the hallmarks of aging that cause decline from those that are consequences, using interventions in cells and animals.
- Safe partial reprogramming Medium AI leverageReprogram old cells to a younger state without losing identity or causing cancer, now at first-in-human stage.
- Clearing old cells safely High AI leverageFind drugs that remove damaged senescent cells selectively without harming healthy tissue.
- Faster aging trials in people Low AI leverageShorten trials of aging treatments with surrogate endpoints, targeted populations or digital twins.
Where AI could help
Medium AI leverage. AI can find targets and build aging clocks, but proving a treatment slows human aging needs years-long trials and accepted endpoints.
- Training and validating biological-age clocks on large cohorts
- Screening for senolytics and rejuvenation targets in cell models
- Designing safer, more efficient reprogramming factors
- Mining biobank and trial data for early biomarkers that predict later decline
Shown so far
- In May 2023 Nature Aging reported that graph neural networks screened over 800,000 molecules and found three drug-like senolytics, one of which cut senescent cell burden in the kidneys of aged mice. source
- On 7 September 2026 Nature Biotechnology published a 43-patient trial in which all six aging clocks showed lower predicted age after 12 weeks on Insilico's AI-generated drug rentosertib, in patients with lung fibrosis, not healthy people. source
- In August 2025 OpenAI and Retro Biosciences said their GPT-4b micro model designed reprogramming-factor variants, with over 30% of SOX2 variants beating the natural protein (according to the companies). source
Prerequisites
- Telomeres and the Hayflick limit1961-1985Cell senescence and shortening telomeres are core ideas in the biology of aging
- Induced pluripotent stem cells2006
- Base and prime editing2016
- Single-cell atlases2016
- GLP-1 weight-loss drugs2021