For most of our lives we’ve been handed the same message about how long we’ll live: eat well, stay active, manage your stress, don’t smoke, and you’ll be rewarded with more years. The implication was always that your choices were the dominant force. Your genes were a footnote.

A new study says that’s probably wrong — and the correction is bigger than most people realize.

What Scientists Just Found

Researchers at the Weizmann Institute of Science in Israel published a study in January in the journal Science that fundamentally revises how much of your lifespan is actually written in your DNA. According to the findings, genetics accounts for about 50% of the variation in human lifespan — twice as much, or more, than previously thought.

To understand why this is a big deal, you need to know what the old number was. Earlier twin studies suggested that only 20–25% of variation in lifespan was genetic, and some large family tree studies put the number closer to 6%. A 2018 study from Alphabet’s longevity research arm Calico pushed the estimate below 10%. For a long time, the scientific consensus was essentially: your DNA doesn’t matter much. Your environment does. That framing shaped decades of research priorities and funding decisions.

The new study says that consensus was built on dirty data.

The Data Problem That Skewed Everything

The original estimates came from studies of twins born between 1870 and 1900 — an era defined by high rates of deaths caused not by biological aging, but by external forces: infectious diseases like tuberculosis and pneumonia, dangerous working conditions, and limited medical care.

The problem was that nobody separated those deaths from the deaths that came from aging itself. A 25-year-old killed in a road accident contributes to mortality statistics in a way that has nothing to do with their genetic predisposition to age-related disease. But the old models counted it anyway — and when one twin died at 32 from typhoid while the other lived to 87, the statistical model recorded it as evidence that genes didn’t matter.

Extrinsic mortality — deaths from accidents, infections, and violence — was about 10 times higher at the time those studies were conducted than it is today. Centuries of noise in the data buried the genetic signal completely.

The Weizmann team corrected for this by building mathematical models that filtered out non-aging deaths, and by doing something no study of this type had done before: they included data from twins who were raised apart, which lets you cleanly separate what genes contribute from what a shared household environment contributes. Once the data was cleaned and the methodology tightened, the genetic contribution to lifespan didn’t nudge upward slightly. It doubled.

As someone who works with data every day, this particular finding resonates at a different level. It’s not that the earlier scientists were wrong — it’s that their data included confounding factors that made it impossible to see the actual signal. The senior author of the study said it plainly: “Those papers weren’t wrong, they just included a lot of confounding factors.” Clean your data first. The findings change when you do.

What the 50% Actually Means

The number doesn’t mean your lifespan is fixed. It means roughly half the variation in how long different people live traces back to their genetic makeup — and the other half is still environment, choices, and chance. Uri Alon, the study’s senior author, describes genetics as a kind of “genetic set point” — a built-in biological influence that shapes but doesn’t entirely determine how long you’ll live.

The more specific numbers are worth sitting with. Up to age 80, the genetic contribution to dementia risk sits at about 70% — substantially higher than for cancer or heart disease. That’s not a rounding error. That’s a finding that should be redirecting research dollars.

And there’s a practical anchor from a longevity researcher at Boston University: even without a favorable genetic hand, the average person can likely reach 88 years for men and 93 years for women through healthy living. So lifestyle still moves the needle meaningfully. It just may not be the whole story.

Personally, none of this changes my commitment to eating well and staying active. But it does change how I think about it. The wellness industry has spent decades selling the idea that the right habits can effectively override your biology — that longevity is something you earn. This study suggests the ceiling on what your choices can accomplish may be lower than we’ve been told. That’s not a reason to give up. It’s a reason to be honest about what we’re actually controlling.

Where AI Comes In — And Why This Gets Interesting

Here’s where the finding connects to something bigger. The reason low genetic estimates mattered wasn’t just academic — they may have discouraged funding and research into the genetics of aging, suggesting the field was largely random or environmental. For decades, researchers poured resources into lifestyle interventions and largely moved on from genetic targets. Now that door is reopened.

And the timing couldn’t be better, because the tools available to walk through it are genuinely different than they were even five years ago. Many labs are now using AI to speed up the drug discovery process — one CEO at Insilico Medicine has 30 AI-based projects underway specifically chosen for their “dual purpose” potential, targeting both specific diseases and the aging process simultaneously.

OpenAI has already built an AI model for longevity science in partnership with Retro Biosciences — a company backed by $180 million from Sam Altman — specifically to engineer better versions of proteins associated with cellular rejuvenation. The goal is extending normal human lifespan by 10 years. That’s not a moonshot fringe project anymore. That’s a well-funded program with real-world lab results.

The broader point is that the Weizmann study isn’t just a finding — it’s a permission slip. If genetics drives roughly half of lifespan variation, there’s now a strong rationale for large-scale studies to identify longevity gene variants, with those discoveries potentially fueling drug development to mimic what naturally long-lived people’s genetics already do. With AI dramatically compressing the timeline from genetic discovery to therapeutic candidate, breakthroughs that might have taken 30 years could arrive significantly faster.

It’s worth watching. The combination of better data methodology, a revised understanding of how much genes actually matter, and AI-powered research acceleration is pointing at something genuinely interesting in aging science over the next decade.

The data was dirty. Now it’s cleaner. And the findings that follow from clean data tend to be the ones that actually change things.

Brandon’s Take:

I like staying active and eating well — and I’m going to keep doing it. But this study does shift something in how I think about it. The wellness industry has always implied that the right habits can more or less override your biology, that longevity is something you earn. This suggests the equation is a bit more balanced than that. Half is still on you. Half might have been decided before you made a single choice. What I find genuinely exciting though isn’t the fatalism angle — it’s what happens when you combine this finding with where AI-driven drug research is heading. The timeline for real breakthroughs in aging science is compressing fast. That’s worth paying attention to.