Application of artificial intelligence in synbiotic and functional food development for precision nutrition

Clinical Snapshot

Key Findings

• P Value: Not applicable — no primary statistical analysis performed

• Effect Size: Not quantified — narrative synthesis only; no pooled effect sizes reported

• Primary Outcome: AI and ML approaches can assist synbiotic and functional food development across five domains: ingredient pairing and formulation optimisation, probiotic viability and shelf-life forecasting, functional and antioxidant bioactivity prediction, sensory and consumer preference modelling, and personalisation for precision nutrition

• Nnt Or Sensitivity: Not applicable — no clinical trial data synthesised; no diagnostic accuracy, NNT, or hazard ratio estimates provided. AI model performance metrics (e.g., R², RMSE, accuracy) are referenced qualitatively but not pooled

• Confidence Interval: Not reported — no meta-analytic data available

Clinical Application

Low for immediate clinical translation. Most AI applications described are computational or pilot-scale. Integration into clinical dietetic practice would require validated, interpretable models with demonstrated clinical endpoints, regulatory approval of AI-assisted nutritional claims, and infrastructure for microbiome profiling at scale — none of which are currently established. In the Australian context, functional foods and synbiotics are regulated by Food Standards Australia New Zealand (FSANZ) under the Australia New Zealand Food Standards Code. Health claims for probiotic and prebiotic products require substantiation under Standard 1.2.7, and therapeutic claims trigger TGA oversight under the Therapeutic Goods Act 1989. The Therapeutic Goods Administration (TGA) does not currently have a regulatory pathway specifically for AI-assisted nutritional product development. The RACGP does not yet have specific guidelines on AI-assisted precision nutrition. The PBS does not list synbiotic or functional food products. Australian dietitians (Dietitians Australia) and gastroenterologists should note that this review does not provide evidence sufficient to alter current clinical practice or support new product health claims under FSANZ standards. The review's findings are relevant to Australian food technology researchers and industry innovators but require substantial clinical validation before influencing patient care pathways. Currently applicable only to food scientists, nutrition researchers, and product development teams working in computational or pilot-scale settings. Not yet applicable to individual patients or clinical populations. Future applicability may extend to individuals with metabolic syndrome, inflammatory bowel conditions, or those requiring microbiome-targeted dietary interventions, but this requires prospective clinical validation.

Abstract

Synbiotic and functional food formulations modulate gut microbiota, SCFA production, immune signalling and metabolic pathways, yet current development pipelines remain largely empirical and constrained by nonlinear trade-offs among probiotic viability, prebiotic functionality and sensory acceptance. Artificial Intelligence (AI) and machine learning (ML) approaches offer data-driven strategies to support formulation, multi-objective optimization and functional assessment. This review systematically identified and critically synthesized literature published between 2020 and 2025 across five thematic domains: (i) formulation and ingredient selection; (ii) viability and shelf-life modelling; (iii) functional and antioxidant bioactivity assessment; (iv) sensory and consumer prediction; and (v) personalization and precision nutrition. Studies were identified through database searches and screened for relevance to synbiotics, functional foods, microbiome modulation and nutrition outcomes, following PRISMA 2020 guidelines. Current evidence indicates that AI can assist ingredient pairing, viability forecasting, sensory modelling and functional property prediction, often complementing conventional statistical tools such as Response Surface Methodology in multivariate design spaces. However, most implementations remain computational or pilot-scale, with minimal integration of microbiome-informed personalization, clinical endpoints or adherence outcomes. Major translational gaps include data heterogeneity, model interpretability, regulatory substantiation and scarcity of longitudinal evidence. AI should therefore be considered a complementary decision-support tool that can accelerate hypothesis generation and formulation refinement rather than substitute mechanistic validation or human trials. Bridging computational modelling with microbiome science and nutritional evidence may enable precision synbiotic strategies and next-generation functional food innovation.

References

1. 1.Janani, B., & Adeyeye, S. A. O. (2026). Application of artificial intelligence in synbiotic and functional food development for precision nutrition. Food & Function. https://doi.org/10.1039/d6fo00295a