Beyond Uptake:

Who Gets to Shape Nutrition Knowledge When Both the Science—and the Technology Carrying It—Are Changing?

WANDA founder Tambra Raye Stevenson will join a National Academies workshop on community-engaged nutrition research to ask a timely question: In an era of AI, algorithms, misinformation, and eroding institutional trust, is getting people to “uptake” nutrition science enough?

By WANDA

A pregnant mother can leave a medical appointment with one recommendation about what to eat, encounter a contradictory claim on TikTok before reaching the parking lot, ask an AI chatbot to settle the disagreement that evening, and wake the next morning to a feed full of products, influencers, and algorithms responding to questions she never knowingly asked.

Her problem is no longer simply access to nutrition information. It is deciding what and whom to trust. That increasingly complicated information environment will form part of the backdrop when public health nutritionist and health communication scholar-practitioner Tambra Raye Stevenson, MPH, MA, founder and CEO of WANDA: Women Advancing Nutrition, Dietetics, and Agriculture and also a PhD Candidate at American University School of Communication, participates in an upcoming National Academies of Sciences, Engineering, and Medicine workshop examining how community-engaged research can improve diet, nutrition, and health.

The two-day virtual workshop, Community Engaged Research to Improve Diet, Nutrition, and Health, will take place September 3–4, 2026, from 11 a.m. to 3:30 p.m. Eastern each day. The National Academies says the workshop will examine how community-engaged research might strengthen the uptake of evidence-based dietary guidance, while exploring how community is defined, the methods and challenges involved in conducting community-engaged research, and the potential for these approaches to improve diet and health.

Stevenson’s presentation, “Beyond Uptake: Trust, Belonging, and the Co-Creation of Nutrition Knowledge,” takes that premise one step further.

“We have spent a great deal of time asking how to get communities to receive, understand, and act on nutrition science,” Stevenson said. “I want us to also ask what happens upstream. Who gets to shape the research question? Whose knowledge is considered legitimate? Who explains uncertainty? Who controls the platform carrying the message? And, increasingly, what happens when an algorithm or AI agent becomes the messenger?”

For Stevenson, those are not questions about technology alone. They are questions about power.

When the Science Changes and So Does the Messenger

Nutrition has always wrestled with changing evidence. Fat moved from villain to a more complicated conversation about type and context. Dietary advice around allergens changed as new research accumulated. Debates over meat, carbohydrates, sodium, supplements, ultra-processed foods, and personalized nutrition continue to evolve alongside the science.

Scientific revision is not evidence that science has failed. It is part of how science works.

But the systems communicating that revision are changing just as quickly.

A generation ago, dietary advice might have moved from researchers to government agencies, health professionals, newspapers, magazines, television programs, schools, and families. Today, that knowledge also travels through search engines, social-media influencers, recommendation algorithms, online grocery platforms, generative AI systems, and increasingly conversational agents capable of producing personalized answers in seconds.

That creates what Stevenson describes as a “double-evolution problem.” “The science is evolving at the same time that the information system carrying the science is evolving,” she said. “So the public is not only being asked to understand why the evidence changed. People increasingly have to determine whether the system interpreting the evidence is trustworthy in the first place.”

The nutrition profession has started grappling with that transition. In April, the Academy of Nutrition and Dietetics and the American Society for Nutrition published a joint artificial intelligence and machine learning resource guide aimed at supporting responsible AI use across nutrition practice and research.

Public health is wrestling with the same questions at a broader institutional level. In June, the Association of Schools and Programs of Public Health released an AI framework focused on responsible implementation across education, research, practice, and policy, emphasizing equity, transparency, accountability, human judgment, and community voice.

And the conceptual landscape is moving even faster. A July 2026 paper in Health Promotion International proposed treating digital determinants of nutrition as a distinct area of research and policy, arguing that digital systems are now shaping nutrition through personalized nutrition tools, digital nutrition education, online food environments, digital marketing, and digitally transformed food systems. In other words, the algorithm is no longer simply carrying the nutrition message. It may increasingly help shape the food environment itself.

From a Food Supply Chain to a Knowledge Supply Chain

Stevenson uses an analogy familiar to anyone who studies food systems. Consider a ground beef patty. Meat may move through farms, processors, distributors, and other supply-chain actors before reaching a plate. Once those inputs are blended into a seemingly uniform product, the finished patty does not reveal its entire history. Traceability systems are needed to reconstruct its provenance.

Nutrition information increasingly presents a similar challenge. An AI-generated response may appear as one fluent answer, but the knowledge underneath it may draw from scientific literature, government recommendations, commercial websites, journalism, health information, cultural assumptions, and older material absorbed during model training. The seams disappear.

Stevenson calls this emerging problem the nutrition knowledge supply chain. “If provenance matters for food, why shouldn’t provenance matter for nutrition knowledge?” she said. “A recommendation can sound incredibly convincing without the person receiving it knowing where the evidence originated, what was excluded, what remains uncertain, who funded the underlying knowledge, or what commercial incentives are shaping how it reached them.”

That concern expands the meaning of health literacy. It also complicates trust. Trust in a dietitian is not identical to trust in a federal agency. Trust in a university is not identical to trust in a social-media platform. And trust in a warm, conversational chatbot is not evidence that the system behind the chatbot is trustworthy.

Researchers are already beginning to develop instruments for understanding how dietitians perceive trust in AI, including dimensions such as data security, professional competence, scientific accuracy, and usability. Stevenson believes community-engaged nutrition research should go further. “Trust is what people give,” she said. “Trustworthiness is what institutions and systems have to demonstrate.”

Belonging Is More Than Seeing Yourself on the Screen

That distinction becomes even more important as AI grows more personalized.

Digital systems can increasingly adjust language, recommend culturally familiar foods, generate avatars that resemble particular communities, and imitate tones of voice associated with familiarity and care. That might make nutrition communication more relevant. It also creates a difficult question: Can a system make someone feel seen before it has earned the right to be trusted?

Stevenson’s doctoral research focuses on belonging, and she is bringing that intellectual lens into nutrition. Her working definition moves belonging beyond representation alone: the experience of being recognized, respected, and able to participate meaningfully in the people, processes, places, and knowledge systems shaping food and health. That distinction carries historical weight.

Food and media companies understood long before the arrival of generative AI that culturally familiar characters could create recognition, comfort, and commercial trust. But familiarity did not necessarily mean that the communities represented possessed authority, ownership, or economic benefit in the systems behind those representations.

The digital era raises a modern version of that question. An AI avatar can have a Black face. A chatbot can speak in culturally familiar language. A nutrition platform can recommend jollof rice, collard greens, fonio, or callaloo. But who owns the technology? Who governed its design? Who decided what counted as authentic? Who gets paid? And who can correct the system when it gets the community wrong?

For community-engaged research, that moves the conversation beyond participation toward governance. Representation is not the same as authority. And cultural likeness is not necessarily cultural ownership.

What Narrative Medicine Can Teach Nutrition

Stevenson is also developing an emerging concept she calls narrative nutrition, inspired in part by narrative medicine and her experience in nutrition education, health communication, and community-based work. Narrative nutrition asks researchers to study not only what people eat, but the stories through which food acquires meaning.

Who taught someone to cook? Which meal represents care? Which food carries shame?

Which foods became associated with pregnancy, grief, migration, celebration, poverty, or home? Which advice came from a grandmother, clinician, influencer, church member, advertisement or chatbot?

A 2024 scoping review of narrative medicine describes the approach as one that brings clinical knowledge into conversation with patients’ lived experiences and uses narrative and interpretive skills to understand perspectives that conventional clinical data can miss.

Stevenson believes nutrition needs a comparable capacity.

“A dietary recall may tell me that a mother did not eat the vegetable we recommended,” she said. “Her story may tell me that the store did not carry it, she could not store it, nobody else in the household would eat it, or it carried a cultural meaning the intervention never bothered to understand. The measure identifies the behavior. The narrative may help us understand the mechanism.”

That approach informs WANDA’s NOURISH Maternal Food as Medicine work, where Stevenson argues that nutrition interventions should begin not only with what pregnant and postpartum women are instructed to consume, but also with questions about family foodways, caregiving, culture, access, competing health information, and what nourishment means to the mother herself.

It also shapes WANDA’s World, the organization’s emerging children’s educational platform, where Little WANDA is imagined not simply as the recipient of nutrition instructions but as a “Food Shero” and “Meal Healer”— a child who investigates food, culture, agriculture, marketing, and health. The distinction matters. People do not consume facts alone. They consume meaning.

Community Engagement in the Age of AI

That is where the National Academies workshop becomes particularly timely. The National Academies describes community-engaged research as a way to help bridge the gap between scientific evidence and meaningful, motivating approaches to healthy eating. But as nutrition research becomes increasingly digital, the question is whether communities will simply be recruited to test new technologies or participate in deciding what those technologies are allowed to become.

Recent research illustrates how far there is to go. A 2026 article on trust and community-engaged health research argues that equal distribution of power, compensation of community partners, sustained partnerships, and institutional support are fundamental to building trustworthy research relationships.

Stevenson wants those principles carried into AI and digital nutrition research.

That means asking communities not only whether they like an interface, but whether they should have decision-making authority over data use, cultural interpretation, research questions, evaluation criteria, intellectual property, ownership, and benefit sharing.

Susan C. Scrimshaw, chair of the National Academies workshop planning committee, said Stevenson brings a perspective particularly suited to that conversation:

“Tambra Raye Stevenson brings a distinctive and timely perspective to this workshop. Her experience across nutrition, public health, communication, community leadership, and food policy allows her to connect scientific evidence with the cultural realities, relationships, and systems that shape how people understand and use nutrition guidance. Her presentation will challenge us to think beyond simply communicating research toward building trustworthy partnerships in which communities help shape the knowledge intended to serve them.”

A New Question for Philanthropy

The changing nutrition information landscape should also matter to philanthropy.

For years, funders have invested in nutrition education, food access, digital health tools, childhood obesity prevention, maternal health, media literacy, and workforce development. The AI era creates an opportunity to connect those funding streams rather than building another generation of isolated interventions.

The philanthropic question is no longer simply whether to fund a new nutrition app, chatbot, video series, or culturally tailored campaign. It is whether to fund the public-interest infrastructure around them.

One emerging model comes from Common Sense Media’s Youth AI Safety Institute, which conducts independent third-party evaluations of AI products used by children and teens. Its methodology goes beyond technical performance to examine developmental appropriateness, fairness, human relationships, misinformation, privacy, transparency, and severe harms. The institute is funded by philanthropy and industry while stating that it retains control over its standards, research, and published evaluations.

Nutrition needs an equivalent conversation.

• Who independently tests a child-facing nutrition chatbot?

• Who evaluates whether an AI system handles eating-disorder risk appropriately?

• Who assesses whether culturally tailored nutrition advice is accurate rather than stereotyped?

• Who studies whether a conversational agent is generating appropriate trust—or merely sounding trustworthy?

• And who ensures that historically underrepresented communities are not only represented in these technologies, but trained, funded, and positioned to research, design, govern, own, and benefit from them?

For philanthropy, that could mean investing in longitudinal research, community-governed studies, independent technology evaluation, fellowships linking nutrition with communication and AI ethics, culturally diverse research datasets, digital nutrition literacy, maternal and child health research, community media, and a stronger pipeline of nutrition professionals prepared for an algorithmic information environment. The opportunity is bigger than “AI for good.” It is to build the human and institutional infrastructure that determines what good means and who has the authority to define it.

Beyond Uptake

Stevenson’s title is intentionally provocative. 'Beyond Uptake' does not mean uptake does not matter. It means uptake may be the wrong place to begin. Before asking why someone failed to follow a recommendation, researchers might ask whether the recommendation was relevant, affordable, understandable, culturally legitimate, supported by the environment, transparently communicated, and worthy of trust.

Before calling a technology inclusive, researchers might ask whether communities shaped the system or merely appeared in it. Before celebrating personalization, they might ask whether personalization is actually person-centered. And before telling people to trust science, institutions might ask whether their own behavior demonstrates trustworthiness.

“The future of nutrition cannot only be about creating more information,” Stevenson said. “We need to understand the systems determining whose knowledge is visible, whose stories count, who controls the technology, and who has the power to correct it. The information system is becoming part of the food system.”

That shift could require new methods, new workforce competencies, new policy frameworks, and new funding priorities. It may also require changing the question that nutrition researchers have asked communities for decades. Not simply: Will you follow the evidence?

But: Will you help us decide how nutrition knowledge should be produced, communicated, governed, and used? That is what it means to move beyond uptake.

Attend the National Academies Workshop

Community Engaged Research to Improve Diet, Nutrition, and Health: A Workshop

Dates: Thursday, September 3 and Friday, September 4, 2026

Time: 11:00 a.m.–3:30 p.m. ET each day

Format: Virtual

The National Academies states that the workshop will bring together a range of perspectives to examine how community-engaged research can support evidence-based dietary guidance and improve diet, nutrition, and health.

Register: https://events.nationalacademies.org/47018

Learn more about the National Academies project: https://www.nationalacademies.org/projects/CHPP-FNA-26-02/about

Tambra Raye Stevenson’s presentation:

Beyond Uptake: Trust, Belonging, and the Co-Creation of Nutrition Knowledge