I Gave My Blood to the AI Death Clock

The reveal came on a September morning with a grinning grim reaper lounging in a doughnut-shaped pool float, under a single number: 74. Death Clock, an AI-powered longevity app launched in 2024, had processed the reporter's health questionnaire and predicted her age at death, along with three likely causes: cardiovascular disease, a sleep disorder, or an unspecified cancer. A few days later, when she entered a mild family history of diabetes, the app revised its estimate downward — to 73.

The app says its AI is trained on 1,200 longevity studies, life insurance actuarial data, and life expectancy tables, generating a prevention plan meant to stretch a user's remaining years. In February, the company added the ability to order blood work: a $69 subscription covers a basic panel measuring cholesterol, blood sugars, inflammation, vitamin D, and blood and organ health, with nine additional tests covering hormones, micronutrients, and thyroid function costing $780 altogether. The reporter opted for the routine panel plus a $29 stress test, drawn at Quest Diagnostics, one of two labs the company has partnered with.

Death Clock sits inside a longevity market projected to reach $8 trillion by 2030, where AI-native health services are multiplying fast. Doctronic, a HIPAA-compliant platform functioning like a personal AI doctor, handles prescription refills without a human physician; Superpower and Lifeforce offer membership-based health insights with clinician-reviewed recommendations; Twin Health claims to build an AI replica of a user's metabolism. Death Clock CEO Brent Franson distances his company from figures like Bryan Johnson, whose 'Don't Die' mantra he calls a dangerous pitch for anyone selling something: "The reality is we're all going to die. It's a question of when."

Skeptics question whether any of it can deliver. Mike Ludkovski, a professor of statistics and applied probability at UC Santa Barbara who researches longevity analysis, argues mortality-forecasting platforms "sell the illusion that individual lifespan is a linear process" that can be hacked, adding that the individual triggers of death are inherently random and will never be truly forecastable. A 2024 Aging Cell study likewise cautioned against treating an aging clock as a measure of overall biological age.

Whether the numbers move — and whether that movement means anything — remains the open question. For builders, the gap between an impressive prediction UI and a validated clinical signal is where regulatory and reputational risk will concentrate as AI health services scale toward that projected $8 trillion market. Watch for how these platforms respond to actuarial critiques like Ludkovski's, whether biomarker-panel offerings draw scrutiny, and whether personalized AI medicine can demonstrate outcomes that survive contact with randomness.