To convince consumers and increase appeal of AI-supported products, brands must make the benefits tangible, prove credibility and reduce perceived risk, said Lim.
Rather than leading with the messaging of “we use AI”, brands should focus on what AI improves for consumers, such as better ingredient selection, formula optimization, improved taste and/or texture, greater cost-effectiveness, or increased transparency.
“Explaining where AI is used in the process and why that adds value helps demystify the technology and builds trust. This matters because acceptance grows when AI is seen as solving real consumer pain points,” Lim told NutraIngredients.
Data from Mintel shows that 48% of consumers in the US believe AI makes it easier to shop for products online.
Based on Lim’s observations, AI is set to reshape the nutraceutical industry across preventative health and speeding up of innovation, but consumer trust and human oversight will be pivotal to its success.
“Over the next five years, AI-enabled nutrition assessment tools are expected to help consumers establish baseline intake, identify nutrient gaps early, and prioritize foundational nutrition, supported by healthcare professionals.”
According to Mintel China Health Tech Market Report 2023, 68% of Chinese consumers feel health solutions based on AI, such as health assessments, are trustworthy.
Although AI is increasingly used to predict ingredient functionality, optimize formulations, shorten R&D timelines, and support faster responses to consumer needs, human expertise remains essential to interpret insights, apply cultural and sensory nuance, and ensure ingredients feel relevant, reliable and desirable to consumers.
Early phase of real-world applications
Despite the growing enthusiasm surrounding AI in nutrition as well as its increasing incorporation into research and practice, its measurable impact remains unclear.
A systematic review evaluated how AI-based systems have been implemented in human nutritional interventions and their influence on health outcomes.
Randomized controlled trials (RCTs) and prospective or retrospective cohort studies published on PubMed, Scopus, Google Scholar, SpringerLink, JMIR, and MDPI were searched from January 2020 to March 2025.
Sixteen studies involving 10,863 participants were included, with most being RCTs targeting metabolic disorders. Several trials reported short-term improvements favoring AI-supported interventions in glycemic control, weight reduction, and symptom severity.
“Our findings demonstrate that while nutrition science and AI have often evolved as separate disciplines, there is a growing body of interdisciplinary work leveraging tools from the technology sector, such as deep learning-based image recognition, wearable-integrated monitoring, and forecasting AI,” the authors wrote.
Notably, the review identified emerging applications of AI but also highlighted the limited robustness of the current evidence base—of the 796 studies initially identified, only 16 met the inclusion criteria.
This mirrors the broader observation that much AI-related nutrition research remains in proof-of-concept or pilot study stages with limited translation into practice. Furthermore, most of the studies were conducted in small samples and showed a significant risk of bias.
These findings carry direct implications for technology developers, healthcare providers, policymakers, and companies in the health and nutrition sectors.
For instance, the review supports the integration of AI-assisted nutrition tools into chronic disease management programs, particularly for obesity, type 2 diabetes, and metabolic syndrome, where their potential to enhance adherence and outcomes is most evident.
At the same time, ethical and regulatory considerations remain crucial.
“The limited number of fully implemented clinically validated tools reflects both the nascency of the regulatory framework and ongoing concerns about data privacy, algorithmic bias, and accountability.
“Premature deployment without rigorous validation risks exacerbating inequities, especially if models are trained on non-representative datasets.”
As such, future research should prioritize multi-center RCTs across diverse demographic and socioeconomic contexts, integration of underutilized data streams such as metabolomics, exposomics and geospatial food-environment mapping, cost-effectiveness analysis, and development of regulatory policies capable of evolving alongside technological innovation.
While the narrative surrounding AI in nutrition is optimistic, this review suggests that its true potential will only be realized through deliberate, evidence-based integration into holistic nutrition care models.
“By combining algorithmic precision with human expertise and embedding AI within supportive behavioral and policy frameworks, the field can move toward equitable, scalable, and impactful nutrition interventions globally,” the authors concluded.
