When concerns about ultra-processed foods started surfacing more publicly, the food industry’s response was consistent:
“We already provide labels.”
On the surface, that sounds reasonable. From a systems perspective, it isn’t.
Nutrition labels were designed to disclose components, not to explain behavior.
They tell you how much sugar, fat, sodium, or protein is present. They do not tell you how a food was engineered, how its ingredients interact with each other, or how it performs as a whole once consumed repeatedly in the real world.
That distinction matters.
In complex systems, more information doesn’t automatically produce understanding. In fact, more partial information can be misleading if it obscures what actually drives [health] outcomes.
A standard nutrition label exposes a small number of attributes – generally 15.
Modern food behavior emerges from far more: processing level, ingredient refinement, additive interactions, texture, consumption speed, and how all of those combine to affect satiety and metabolism over time. None of that fits cleanly on a label.
This isn’t a failure of effort. It’s a limitation of the model.
Labels assume that if people are given enough numbers, they can infer the rest. But inference breaks down when food systems become too complex. The burden quietly shifts to the consumer to interpret effects that even experts struggle to model without proper data.
That’s why adding more labels never solved the problem.
It was a lack of interpretability.
In other engineering domains, we don’t hand users raw telemetry and call it transparency. We summarize, contextualize, and surface system-level behavior. Food labeling, historically, stopped at the component level because that was once sufficient, but now it isn’t anymore.
What’s changed is not consumer intelligence. It’s system complexity, and our ability to measure it.
Once processing levels, formulation patterns, and population-scale exposure can be quantified, the limits of traditional labels become obvious. They weren’t wrong. They were simply designed for a simpler world.
Labels didn’t fail because they hid information.
They failed because they couldn’t explain how the system behaves.