AI FAQ
How and Why We Use AI
We use AI in two key ways.
Developing the Standard — We used a custom AI tool to pressure-test and refine the requirements in the Non-UPF Verified Standard. This helped us find the right balance between meaningfulness (delivering a genuine benefit for shoppers) and achievability (ensuring the requirements were realistic for manufacturers). That balance is foundational to the program's ability to shift the food supply toward minimally processed foods.
Pre-verification screening — Now that the program is in open enrollment, AI serves as a pre-verification screening tool. Brands curious about the program can complete a basic assessment using on-package information to get a clearer picture of what documentation they'd need — or whether any ingredients in their formulation may need to be replaced — before making a significant investment of time or money.
AI is a valuable tool, but it has limitations — and being transparent about both AI use and its limits is an important step toward rebuilding trust with shoppers and eaters.
AI can't evaluate manufacturing intent. The same ingredient can serve very different purposes: a gum might be used to ensure even nutrient distribution in a fortified beverage, or it might be used to mimic the creaminess of a whole food with no nutritional benefit. AI sees the ingredient, not the "why" behind it. That distinction is central to UPF assessment, and it requires human judgment.
AI has no sensory experience. Product formulation is ultimately a human endeavor. AI can flag parts of a formulation that may not align with the Standard, but it can't guide the process of fixing it. Ultimately, food is created, shared and enjoyed by people.
AI's benefits depend on the surrounding system. Pre-verification screening only improves accessibility and transparency when the Standard is publicly available and when the process is guided by human relationships.
AI is a useful starting point in certification, not a substitute for human knowledge and expertise.
For brands, pre-verification screening helps smooth the pathway to verification. Brands arrive better prepared, which makes the human review process more focused and efficient — a better use of everyone's time and resources.
For shoppers, the benefit is more Non-UPF Verified products on shelf and, ultimately, a market shift toward less processed foods. Using AI-assisted tools early in the process helps brands efficiently meet a rigorous standard — and that balance of rigor and accessibility is what drives real change.
For more information, see the Non-UPF Verified Brand Starter Guide.
Before enrolling in the program, brands can complete an AI-assisted pre-screening with our team. This helps identify any formulation areas that may need extra attention before verification, so you arrive prepared for the process.
Yes. The AI tools we use are designed by people, and the results are assessed and communicated to brands by people. Pre-verification screening results are reviewed by our team before any feedback reaches a brand. Those interactions also help us refine the rules and parameters we give the AI tool over time — improving its accuracy and deepening our own understanding of the non-UPF landscape.
The AI tools used in pre-verification screening analyze on-package information: product name, brand name, ingredient list and nutritional information. This gives Technical Administrators an informed starting point, freeing them to focus their expertise where it matters most — assessing processing methods and intensity, which requires human judgment and experience.
No. A rigorous certification requires reviewing manufacturing documentation, ingredient sourcing and production practices, not just label data. Without access to that information, we don't believe it is possible to credibly verify whether a food is ultraprocessed. AI contributes to the process; it doesn't determine the outcome.
Our mission is to empower people to care for themselves, the planet, and future generations. Those commitments, and our core belief that food should nourish life, shape how we approach every tool we use, including AI.
We use AI to augment human work, not replace it; to deepen our understanding and grow our capacity to make positive change. The goal is to make our team's expertise more effective and our program more accessible to brands, not to automate or cut corners. Where AI analyzes and contributes, humans ideate and decide.
We're also clear-eyed about the fact that AI is not a neutral technology. Its development and environmental footprint have significant impacts — as does its potential to either enhance or diminish human potential. That's why we think carefully about where AI adds genuine value and where it doesn't belong. We are committed to continually evaluating our use of AI as the technology, and our understanding of it, evolves.
