Abstract Objective To assess the effectiveness of an artificial intelligence (AI)-designed personalized diet compared to a standard Mediterranean diet in patients with metabolic risk factors over a 12-week period. Materials and Methods This randomized controlled trial (RCT) included slightly more than 40 patients with overweight (BMI 25–35 kg/m²) and metabolic risk factors. Participants were randomly assigned in a 1:1 ratio to either an AI-personalized diet group or a standard Mediterranean diet group. The AI diet was developed using machine learning algorithms that analyzed individual anthropometric, biochemical, and behavioral parameters to tailor macronutrient and micronutrient intake dynamically. Both groups received regular dietary consultations and maintained food diaries. Results At the end of 12 weeks, both groups demonstrated positive metabolic changes, with the AI-based diet group achieving more pronounced, yet physiologically reasonable improvements: • Weight Reduction: The AI diet group lost an average of 3.1 ± 0.9 kg, compared to 1.9 ± 0.7 kg in the control group (p = 0.04). • BMI and Waist Circumference: BMI decreased by 0.9 kg/m² (vs. 0.5 kg/m², p = 0.05), and waist circumference reduction was 5.2 ± 1.3 cm (vs. 3.1 ± 1.0 cm, p = 0.03). • Lipid Profile: The AI group showed a reduction in LDL cholesterol (15.8% vs. 9.5%, p = 0.04) and triglycerides (18.3% vs. 10.9%, p = 0.05), along with a modest increase in HDL cholesterol (4.2% vs. 2.7%, p = 0.06). • Glycemic Control: Fasting glucose levels improved slightly in both groups, with a reduction of 0.42 ± 0.08 mmol/L in the AI group versus 0.25 ± 0.07 mmol/L in the control group (p = 0.05). The insulin resistance index (HOMA-IR) improved by 8.7% in the AI group versus 4.9% in the control group (p = 0.07). • Subjective Health Parameters: Patients following the AI diet reported improved sleep quality (PSQI score reduction of 12% vs. 7%, p = 0.08), increased energy levels (+9% vs. +5%, p = 0.06), and slightly reduced digestive discomfort (56% vs. 41%, p = 0.09). Conclusion The AI-driven personalized diet demonstrated slightly superior effectiveness in weight management and metabolic health improvement compared to the standard Mediterranean diet. While the changes remained within expected physiological limits, the AI approach showed potential in optimizing dietary personalization and patient adherence. Further long-term studies are needed to validate these findings and assess sustainability beyond the 12-week intervention.
Rozibaev et al. (Sat,) studied this question.