Forecast: By 2045, AI will be better at diagnosing rare diseases than the best human specialist. Average life expectancy in Switzerland will rise to over 95 years.
Disease is in large part an information problem: the body sends signals that are too weak, too complex or too numerous for a human doctor to fully interpret. AI can read these signals.
Already today, AI systems diagnose certain types of skin cancer with greater accuracy than experienced dermatologists [1]. DeepMind's AlphaFold solved the protein structure prediction problem -- an issue biologists had worked on for fifty years -- in months [2]. These are no longer laboratory experiments. They are tools transitioning into clinical practice.
Scenario for 2040: Continuous health monitoring via wearable sensors will be standard. Heart rhythm, blood pressure, blood sugar, inflammation markers, sleep quality -- all captured in real time, analysed by AI, compared with millions of anonymised patient records.
The University Hospital Zurich and the CHUV Lausanne are already cooperating with ETH and EPFL to integrate AI into clinical diagnostics [3].
Stem cells from the patient, frozen during youth, make it possible to replace damaged organs with tailor-made spare parts -- without rejection, without lifelong immunosuppression. What is considered experimental medicine today will become routine therapy.
The heart that fails after decades receives a successor grown from the patient's own tissue. First clinical trials with tissue grown from stem cells are already underway in Japan and the USA [4].
Today's medicine treats diseases according to average values: one drug, one dose, for everyone. Personalised medicine reverses this principle. AI analyses millions of patient records, genetic profiles and treatment outcomes and identifies which therapy works best for which patient.
Forecast: Life expectancy in Switzerland, currently at 84 years, will rise to over 95 -- not in physical decline but in sustained health [5].
The big question: who benefits? The costs of basic medical care will fall dramatically because AI diagnostics are cheaper than specialist consultations. But the main beneficiaries could be the companies that own the medical AI systems -- unless political regulation intervenes.
Switzerland faces a fundamental decision here: will AI-powered medicine be a public good or a premium service? The answer will determine whether the medical revolution benefits everyone or exacerbates inequality.
[1] Esteva, Andre et al.: Dermatologist-level classification of skin cancer with deep neural networks. Nature 542, 2017, pp. 115--118.
[2] Jumper, John et al.: Highly accurate protein structure prediction with AlphaFold. Nature 596, 2021, pp. 583--589.
[3] ETH AI Center / University Hospital Zurich: AI in Clinical Diagnostics -- Cooperation Programme. ai.ethz.ch, 2024.
[4] Yamanaka, Shinya: Induced Pluripotent Stem Cells: Past, Present, and Future. Cell Stem Cell 10(6), 2012, pp. 678--684.
[5] Federal Statistical Office (FSO): Life Expectancy in Switzerland. Neuchatel, 2024. (Current value: 84.0 years.)