Harnessing Generative Artificial Intelligence for Exercise and Training Prescription: Applications and Implications in Sports and Physical Activity—A Systematic Literature Review

Applied Sciences, Journal Year: 2025, Volume and Issue: 15(7), P. 3497 - 3497

Published: March 22, 2025

Regular physical activity plays a critical role in health promotion and athletic performance, necessitating personalized exercise training prescriptions. While traditional methods rely on expert assessments, artificial intelligence (AI), particularly generative AI models such as ChatGPT Google Gemini, has emerged potential tool for enhancing personalization scalability recommendations. However, the applicability, reliability, adaptability of AI-generated prescriptions remain underexplored. A comprehensive search was performed using UnoPerTutto metadatabase, identifying 2891 records. After duplicate removal (1619 records) screening, 61 full-text reports were assessed eligibility, resulting inclusion 10 studies. The studies varied methodology, including qualitative mixed-methods approaches, quasi-experimental designs, randomized controlled trial (RCT). ChatGPT-4, ChatGPT-3.5, Gemini evaluated across different contexts, strength training, rehabilitation, cardiovascular exercise, general fitness programs. Findings indicate that programs generally adhere to established guidelines but often lack specificity, progression, real-time physiological feedback. recommendations found emphasize safety broad making them useful guidance less effective high-performance training. GPT-4 demonstrated superior performance generating structured resistance compared older models, yet limitations individualization contextual adaptation persisted. appraisal METRICS checklist revealed inconsistencies study quality, regarding prompt model transparency, evaluation frameworks. holds promise democratizing access prescriptions, its remains complementary rather than substitutive guidance. Future research should prioritize adaptability, integration with monitoring, improved AI-human collaboration enhance precision effectiveness AI-driven

Language: Английский

---

Assessment of recommendations provided to athletes regarding sleep education by GPT-4o and Google Gemini (Preprint)

Published: Jan. 16, 2025

Inadequate sleep is prevalent among athletes, affecting adaption to training and performance. While education on factors influencing can improve behaviors, Large Language Models (LLMs) may offer a scalable approach provide athletes. This study aims i) investigate the quality of recommendations generated by publicly available LLMs, as evaluated experienced raters, ii) determine whether varies with information input granularity. Two prompts differing granularity (low high) were created for two use cases inserted into ChatGPT-4o (GPT-4o) Google Gemini, resulting in n=8 different recommendations. Experienced raters (n=13) 1-5 Likert-scale, based n=10 criteria derived from recent literature. The highest summary rating was achieved GPT-4o using high granularity, ratings >3 (tendency towards good), n=3 equal 3 (neutral), n=2 <3 bad). significantly outperformed Gemini 9 out 10 (P<.001 P=.045). Recommendations received higher than those low across both LLMs (P<0.001 P=.049). Both exhibit limitations, neglecting vital education. Sleep suboptimal, achieving overall quality. However, demonstrated improved recommendation emphasizing need specificity thorough review outputs securely implement AI technologies

Language: Английский

---

Evaluating the Potential Role of AI Chatbots in Designing Personalized Exercise Programs for Weight Management

International Journal of Human-Computer Interaction, Journal Year: 2025, Volume and Issue: unknown, P. 1 - 8

Published: Feb. 27, 2025

This study aimed to evaluate the effectiveness and potential use of artificial intelligence (AI) chatbots in developing personalized exercise programs for weight management. Exercise were developed by ChatGPT-4, ChatGPT-4o, Gemini-1.5 Pro models, a group human expert trainers hypothetical obese individual case. All assessed based on American College Sports Medicine (ACSM) National Academy (NASM) guidelines. The chatbot-generated consistent with ACSM NASM standards, indicating their low-resource settings. Nevertheless, considerable differences found between key parameters, including initial load target heart rate zone recommendations. While AI have enhance accessibility, expertise remains essential ensure program safety effectiveness. results this provide insights into role

Language: Английский

---

Assessment of recommendations provided to athletes regarding sleep education by GPT-4o and Google Gemini: A comparative evaluation study (Preprint)

JMIR Formative Research, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 16, 2025

Language: Английский

---

Acceptance and trust in AI-generated exercise plans among recreational athletes and quality evaluation by experienced coaches: a pilot study

BMC Research Notes, Journal Year: 2025, Volume and Issue: 18(1)

Published: March 13, 2025

Abstract Objectives Large language models are becoming increasingly significant tools in everyday life, including the context of training and sports. However, extent to which recreational athletes actually rely on AI-generated plans differences trust towards these technologies between users non-users have not yet been investigated. Furthermore, there is a lack information regarding current quality such plans. The aim this project was examine how differ their assess Results In our sample, 54% participants trained using structured plan, with 25% those utilizing Users AI-based exhibited significantly ( p = 0.030) higher levels compared non-users. output from large has now reached level where even professional coaches often unable distinguish whether plan or created by human expert. This suggests that could potentially match standards developed experienced coaches, making them viable option for seeking guidance training.

Language: Английский

---

ChatGPT-4o-Generated Exercise Plans for Patients with Type 2 Diabetes Mellitus—Assessment of Their Safety and Other Quality Criteria by Coaching Experts

Sports, Journal Year: 2025, Volume and Issue: 13(4), P. 92 - 92

Published: March 24, 2025

In this discussion paper based on preliminary data, the safety and other quality criteria of ChatGPT-4o-generated exercise plans for patients with type 2 diabetes mellitus (T2DM) are evaluated. The study team created three fictional patient profiles varying in sex, age, body mass index, secondary diseases/complications, medication, self-rated physical fitness, weekly routine personal preferences. Three distinct prompts were used to generate each patient. While Prompt 1 was very simple, 3 included more detailed requests. optimized by ChatGPT itself. coaching experts reviewed discussed their evaluations. Some showed serious issues, especially diseases/complications. most incorporated key training principles, they some deficits, e.g., insufficient feasibility. use (Prompt 3) tended result elaborate better ratings. may have issues T2DM, indicating need consult a professional coach feedback before starting program.

Language: Английский

---

Selective Trust: Understanding Human-AI Partnerships in Personal Health Decision-Making Process

Published: April 24, 2025

Language: Английский

---

The Last JITAI? Exploring Large Language Models for Issuing Just-in-Time Adaptive Interventions: Fostering Physical Activity in a Prospective Cardiac Rehabilitation Setting

Published: April 25, 2025

Language: Английский

---

An Outlook for AI Innovation in Multimodal Communication Research

Lecture notes in computer science, Journal Year: 2024, Volume and Issue: unknown, P. 182 - 234

Published: Jan. 1, 2024

Language: Английский

---

What is the potential of ChatGPT for qualified patient information?

Zeitschrift für Rheumatologie, Journal Year: 2024, Volume and Issue: unknown

Published: July 10, 2024

Zusammenfassung Einführung Der Chatbot ChatGPT stellt einen Meilenstein in der Interaktion zwischen Menschen und großen, über das Internet zugänglichen Datenbanken dar. Er ermöglicht mit einer Kommunikation Alltagssprache die Beantwortung komplexer Fragen ist damit potenziell eine Informationsquelle für Betroffene rheumatischer Erkrankungen. Ziel Untersuchung war es herauszufinden, ob (Version 3.5) Lage ist, qualifizierte Antworten zur Anwendbarkeit von Verfahren Komplementär- Alternativmedizin (CAM; Homöopathie, Ayurveda, Phytotherapie) bei rheumatoider Arthritis (RA), systemischem Lupus erythematodes (SLE) Granulomatose Polyangiitis (GPA) zu liefern. Außerdem wurde untersucht, welchen Einfluss Art Fragestellung auf erhaltenen Ergebnisse haben könnte. Methodik Die Befragung erfolgte 3 Abschnitten. In Abschnitt A offene Frage Behandlungsmöglichkeiten einem Krankheitsbilder gestellt. B allgemein nach möglichen Anwendungen CAM Erkrankungen gefragt. C wurden Applikationsmöglichkeiten genannten jede Diagnose erfragt. den Abschnitten jeweils 2 Modifikationen erste fragte danach, überhaupt anwendbar ist. zweite erkundigte sich konkreten aus Verfahren. Validität anhand des Reliability Scores, 7‑stufigen Likert-Skala, ausgewertet. Zu offenen im lieferte validesten Ergebnisse. zahlreiche CAM-Anwendungen vorgeschlagen, nicht durch wissenschaftliche Evidenz gestützt sind. diesen waren deutlich abhängig. Suggerierte Anwendungsabsicht CAM, entfielen häufig Hinweise fehlende Evidenz, Qualität Antwort meisten Fällen schlechter bewertet. Schlussfolgerung Anwendung definierten rheumatischen lassen ausreichende vermissen. Zudem beeinflusst Aussagen erheblich. Eine kritiklose als Instrument Patientenschulung kann derzeit empfohlen werden.

---