Optimizing Weight Loss with Artificial Intelligence: A Randomized Controlled Trial of Dietary-Supplement Prescriptions in Adults with Overweight and Obesity.

Abstract

Background: Obesity is a complex, multifactorial chronic disease that poses significant health risks. Recent advancements in artificial intelligence (AI) offer the potential for more personalized and effective dietary-supplement (DS) regimens to promote weight loss. This randomized controlled trial evaluated the efficacy of AI-guided DS prescriptions compared with standard physician-guided DS prescriptions in adults with obesity.

Methods: This randomized, parallel-group pilot study enrolled 60 individuals aged 40-60 years with a body-mass index (BMI) ≥ 25 kg m^-2. Participants were randomized to receive either AI-guided DS prescriptions generated on the basis of each patient's individualized genetic, metabolic, and behavioral data (n = 30) or physician-guided DS prescriptions (n = 30) for 180 days. The primary endpoints were the percentage change in body weight and the proportion of participants achieving a ≥ 5 % weight reduction. Secondary endpoints included changes in BMI, fat mass, visceral-fat rating, systolic and diastolic blood pressure, lipid profiles, fasting plasma glucose, hsCRP levels, and postprandial appetite ratings. Adverse events were monitored throughout the study.

Results: Baseline characteristics were well balanced between groups. Mean weight loss was -12.3 % (95 % CI: -13.1 to -11.5) in the AI-guided group vs. -7.2 % (95 % CI: -8.1 to -6.3) in the physician-guided group, giving a treatment difference of -5.1 % (95 % CI: -6.4 to -3.8; p < 0.01). At day 180, 25/30 (83.3 %) AI-guided participants achieved ≥ 5 % weight reduction compared with 16/30 (53.3 %) in the physician-guided arm (OR 4.4; 95 % CI: 1.3 to 14.5; p = 0.01). Significant improvements were also seen in BMI, fat mass and visceral-fat rating in the AI-guided group (p < 0.01 for all). Postprandial appetite suppression was greater in the AI-guided group, with significant reductions in hunger and prospective food consumption and increases in fullness and satiety (p < 0.01 for all). Adverse events were generally mild to moderate, with higher incidences of gastrointestinal symptoms in the AI-guided group, but these were manageable and did not affect adherence.

Conclusion: The AI-guided dietary-supplement regimen was more effective in promoting weight loss, improving body composition, and suppressing appetite than the physician-guided regimen. These findings suggest that AI-guided, personalized supplement prescriptions-grounded in genetic, metabolic, and behavioral profiling-could provide a more effective approach to obesity management. Larger studies are warranted to confirm these results and further refine AI-based interventions for weight loss.

Trial registration: ClinicalTrials.gov NCT06458296.