When investigators used artificial intelligence (AI) using a method called Machine-Assisted Topic Analysis (MATA), which combines AI-powered topic modelling with human qualitative analysis, to help them analyze 58,881 X posts referring to the 5 most profitable fitness apps, they identified several negative themes - for example, challenges of quantifying diet and physical activity, the complexity of tracking calories and exercise through oversimplified algorithms, technical challenges and malfunctions, and aversive emotional responses to the apps’ notifications.
As a result, some users noted feelings of shame, disappointment and demotivation, and subsequent disengagement with apps and health behaviors.
The findings highlight the need for a more user-centered and psychologically informed app design that prioritizes wellbeing and intrinsic motivation over rigid, quantitative goals.
"Fitness apps remain some of the most profitable and widely downloaded health tools globally. While they can benefit health, there’s been far less attention to their potential downsides," said corresponding author Paulina Bondaronek, PhD, of University College London. "When health is reduced to calorie counts and step goals, it can leave people feeling demotivated, ashamed, and disconnected from what truly drives lasting wellbeing. Using AI alongside human analysis, we were able to shed light on these often-overlooked impacts by centering the voices of real users."
Sheen F, Porter L, Papakonstantinou T, Ceka M, Bondaronek P.
Living well? The unintended consequences of highly popular commercial fitness apps through social listening using Machine-Assisted Topic Analysis: Evidence from X.
Br J Health Psychol. 2025 Nov;30(4):e70026. doi: 10.1111/bjhp.70026