More than 1,650 researchers, clinicians, and federal regulators will gather this weekend in National Harbor, Maryland, to address a persistent issue in nutrition science: tools designed to guide dietary choices often fail for diverse populations.
The American Society for Nutrition’s annual meeting, NUTRITION 2026, begins Saturday with a full schedule of scientific presentations. One session will highlight a significant problem in AI-powered diet apps. These tools, widely used by clinicians and consumers, frequently miscalculate nutritional content for non-Western cuisines. The error occurs not because the technology is flawed, but because it lacks training on relevant data. That documented limitation, confirmed in peer-reviewed research published just this year, will be among the live questions at the meeting.
The stakes extend beyond simple miscalculations. As AI diet apps enter clinical use, their limitations may disproportionately impact patients who already face care barriers.
Related: Alaska Airlines passenger denied boarding twice over paperwork
The oversight risks widening health disparities, particularly for communities whose dietary patterns have been historically overlooked.
The conference occurs amid heightened scrutiny of nutrition science. The 2025–2030 Dietary Guidelines for Americans were released in January 2026 under conditions of significant scientific controversy, and the NUTRITION 2026 plenary session on Saturday afternoon will mark the first major organized gathering of the scientific community to formally examine what the evidence supports—and where the process went wrong.
Other sessions will cover pressing topics. A Sunday panel on GLP-1 weight-loss drugs will address nutritional complications from appetite suppression, such as protein deficiency and micronutrient gaps. About half of patients stop taking these medications within a year, raising concerns about managing the transition. Meanwhile, a Tuesday session on infant formula regulation will highlight Operation Stork Speed, the federal initiative updating standards that have remained unchanged since the 1980s. Consumer Reports tests found potentially harmful levels of arsenic and lead in some formulas. The U.S. still lacks enforceable limits for heavy metals, though the FDA is working to close this gap.
Related: Chile selects China’s undersea cable over U.S. objections
The meeting’s location near Washington, D.C., was chosen for its proximity to policymakers. With regulatory changes accelerating, the conference provides a rare opportunity for direct engagement. “Connect with the Fed” sessions will allow attendees to interact with leaders from the FDA, USDA, NIH, and CDC. These officials handle challenges ranging from the contested Dietary Guidelines to stricter infant formula standards.
Despite the focus on policy and technology, the conference will revisit fundamental questions. Sessions on protein needs, gut microbiome research, and nutritional drivers of cognitive decline will emphasize the field’s enduring complexity. What people eat, how it affects health, and who defines “healthy” remain difficult to answer.
Thursday Preview: AI and Precision Nutrition Workshop
Before the main conference opens, a pre-conference workshop on Thursday afternoon will introduce attendees to the engineering architecture behind the federal government’s most ambitious AI nutrition research program. The session, titled “AI and Precision Nutrition: Introduction to the All of Us Researcher Workbench,” is co-sponsored by the NIH Office of Nutrition Research, ASN, the Academic Nutrition Departments and Programs (ANDP), and the Cornell Joan Klein Jacobs Center for Precision Nutrition and Health, and will be chaired by Dr. Saurabh Mehta, MBBS, ScD, Director of the Cornell center. Participants will learn to handle the NIH All of Us Researcher Workbench, the cloud-based data platform into which the Nutrition for Precision Health (NPH) study is loading its results. The NPH program, launched with $170 million in NIH funding, is enrolling 10,000 participants from diverse backgrounds to generate AI algorithms that can predict how individual people respond to specific foods—linking genetics, electronic health records, wearable sensor data, and dietary surveys into a unified prediction model. The Workbench itself is hosted on a cloud architecture developed by the US Military Academy’s AIDE-ML Center at West Point, which built the “Data Distiller for Precision Nutrition”—a pipeline that collapses multi-stream data into AI-ready datasets. Participants will need a working knowledge of R and must bring a laptop.