Kan AI generere personlige trænings- og ernæringsplaner, der tilpasser sig i realtid til biomedicinsk feedback ?

AI-drevne fitnessplatforme opretter nu og justerer dynamisk trænings- og kostplaner baseret på live-data fra wearables, pulsmålere og endda stressniveauer. Disse systemer personaliserer anbefalinger ved at analysere søvnkvalitet, restitutionsmålinger og præstationsudvikling. Nogle platforme inddrager genetiske data eller mikrobiomanalyser for at skræddersy ernæringsråd. AI’en lærer af brugerens vaner og justerer intensitet, varighed og kostforslag derefter.

Background

AI-powered fitness platforms now create and dynamically adjust exercise and diet plans based on live data from wearables, heart rate monitors, and even stress levels. These systems personalize recommendations by analyzing sleep quality, recovery metrics, and performance trends. Some platforms incorporate genetic data or microbiome analysis to tailor nutritional advice. The AI learns from the user’s habits and adjusts intensity, duration, and dietary suggestions accordingly.

Current AI systems can generate basic personalized workout and nutrition plans from user inputs such as age, weight, fitness goals, and dietary preferences, and some platforms use static biometric data like heart rate or step count to adjust recommendations. Early-stage research prototypes using wearable streams (ECG, SpO2, temperature, accelerometry) have demonstrated real-time adaptation in controlled lab settings, but these systems remain at feasibility-level rather than clinical-grade reliability, with errors in plan switching when sensor noise or user-context misclassification occurs. Regulatory-approved, real-time closed-loop plans for general use are not yet available. FDA-cleared “digital therapeutic” apps can adapt insulin dosing for diabetics and deliver guided exercise prescriptions, but these adaptations are based on prior-trained models rather than open-loop continuous personalization.

— Enriched May 12, 2026 · Source: U.S. Food and Drug Administration