Three scientists say the trials don't prove processing itself is the problem. Here is the argument without the spin — theirs or ours.

A Perspective in Science today, by nutrition researchers at Copenhagen, Wageningen and Liverpool (DOI 10.1126/science.aef3495), reviews the five controlled feeding trials that compare ultra-processed and minimally processed diets and concludes that they provide weak support for an effect of processing itself on body weight. The authors do not dispute that people ate more and gained more on the ultra-processed diets — in the two inpatient trials, 500 to 800 extra calories a day. Their point is that the ultra-processed diets were also denser in calories, lower in fiber, higher in sodium and faster to eat, and that those properties can plausibly explain all of the difference without invoking "processing" as a separate cause. Four of the five trials lasted one to three weeks; the longest, eight. The authors say short trials cannot rule out slow-accumulating harm, and recommend targeting foods that are nutritionally poor, calorie-dense and rapidly consumed regardless of how processed they are.

Two things a reader deserves to know. First, the argument is a fair one, and independent scientists make versions of it; we have said on this site that whether processing per se harms you beyond what the food carries is not established. Second, the disclosures: the article states that no outside body was involved in writing it and that none of the authors has taken personal fees from the food industry in a decade, and it lists research funding to the authors from dairy, sweetener and confectionery companies and, in the past, a beverage trade association. We report both halves because our rule is to report both halves. A funded author can be right. A reader is entitled to know the funding before deciding.

Loree’s take

Here is what I notice: the authors' practical conclusion and mine are the same. Avoid foods that are calorie-dense, low in fiber, high in salt and designed to be eaten fast. That is a description of the ultra-processed aisle whether or not "processing" is the mechanism. The scientific question — is it the processing or the properties processing produces — matters for policy and labeling. For the person holding the bag of chips, it does not change the answer.

Ben’s take

Read the disclosures the way you would read a cap table. This piece is not industry propaganda; the authors are serious and their disclosures are complete, which is more than most. But notice how it will be used. Within a week, trade groups will cite "a paper in Science" as proof the ultra-processed category is unscientific, and they will not quote the sentence recommending that calorie-dense, fast-eaten foods be targeted anyway. The paper is honest. Its afterlife will not be.

Why it matters to a recovering optimizer. The strongest version of this site beats its best counterarguments on the page, and this is one of them: the controlled-trial evidence for processing as the villain is thinner than the headlines suggest. What the counterargument does not do is rescue the product. By the authors' own standard — nutritionally poor, calorie-dense, rapidly consumed — the box still loses to the kitchen. Eat real food. The mechanism can be argued later.