AINeutralarXiv – CS AI · 14h ago6/10
🧠
Consistency of AI-Generated Exercise Prescriptions: A Repeated Generation Study Using a Large Language Model
A study evaluating the consistency of exercise prescriptions generated by Gemini 2.5 Flash found high semantic consistency but significant variability in quantitative components like exercise intensity. The research highlights that while LLMs produce semantically similar outputs, structural constraints and expert validation are necessary before clinical deployment.
🧠 Gemini