# Proof Narrative: All ultra-processed foods are inherently unhealthy and should be avoided completely.

## Verdict

**Verdict: DISPROVED**

The claim that every ultra-processed food is inherently harmful and that everyone should avoid them entirely is not supported by the major health institutions that study this question — in fact, they say the opposite.

## What was claimed?

The claim is that ultra-processed foods — a broad category defined by industrial production methods and added ingredients — are bad for you by their very nature, and that no one should ever eat them. You've probably seen versions of this online: lists of "foods to never eat," warnings that anything with more than five ingredients is toxic, or blanket advice to cut out all processed foods entirely. The claim matters because ultra-processed foods make up a large portion of modern diets worldwide, and sweeping avoidance advice, if wrong, could create unnecessary anxiety or lead people to reject foods that are actually fine.

## What did we find?

The claim is actually two claims packaged together: first, that every ultra-processed food is inherently harmful; and second, that health authorities recommend avoiding them completely. Both parts are false.

On the first part, the UK's National Health Service explicitly lists ultra-processed foods — wholemeal sliced bread, wholegrain breakfast cereals, baked beans — that it considers acceptable in a healthy diet. These foods are classified as "ultra-processed" under the standard food classification system due to their industrial production, yet the NHS recommends them without hesitation. Harvard's nutrition experts similarly note that some of these products "may be a useful addition to a healthful diet." A peer-reviewed nutrition journal article puts it plainly: "there are good and bad diets, not good and bad foods."

On the second part, not one major health authority — not the WHO, not the NHS, not Harvard — tells people to avoid all ultra-processed foods completely. The WHO says to "limit" foods high in unhealthy fats, free sugars, and sodium. The NHS says most people "would probably benefit from eating less" of certain ultra-processed foods. Harvard frames it as a consumer choice with trade-offs. "Eat less of some" is a very different message from "avoid all."

There is also a structural problem with the claim. The category of "ultra-processed" foods is remarkably wide. Under the standard classification system, it includes diet sodas and commercially produced infant formula in the same group. The American Academy of Pediatrics recommends infant formula for babies who cannot breastfeed — a direct contradiction of any blanket avoidance rule. Even the researchers who created the ultra-processed food classification have acknowledged that no studies have examined what a strict avoidance policy would mean for infants.

The strongest research on this topic — a large Spanish study published in the BMJ — found that people eating four or more servings of ultra-processed food per day had higher mortality than those eating fewer. But the lowest-consumption group, who still ate some ultra-processed foods, had the best outcomes. The harm appears at high intake, not at any consumption whatsoever, and the study's authors were careful to note they couldn't rule out other contributing factors.

## What should you keep in mind?

The disproof of this specific claim does not mean ultra-processed foods are harmless or that dietary patterns don't matter. The evidence consistently supports reducing intake of ultra-processed foods that are high in saturated fat, added sugar, and sodium — chips, sugary drinks, fast food. That advice is well-grounded. What is not grounded is the leap from "eat less of the bad ones" to "all of them are inherently harmful and none should ever be consumed."

The "ultra-processed" label covers an enormous range of products, which makes blanket health claims about the category unreliable. Wholemeal bread and diet soda are not nutritionally equivalent just because both get the same label. The classification system was designed for population-level research and policy, not as a per-product health verdict.

It is also worth noting that the observational research in this area, while suggestive, cannot establish that ultra-processing itself causes harm independent of the nutrients involved. People who eat more ultra-processed foods tend to have different overall diets and lifestyles — separating the effect of processing from the effect of eating more sugar and saturated fat is genuinely difficult.

## How was this verified?

This claim was broken into its two component parts and each was tested against independent authoritative sources, with adversarial checks to look for evidence that could overturn the finding. All six sources were verified by live fetch with direct quote confirmation. You can read the full reasoning in [the structured proof report](proof.md), examine every source and verification step in [the full verification audit](proof_audit.md), or [re-run the proof yourself](proof.py).