The science is messy.

The UPF definition includes a ton of very distinct things. It is unlikely that emulsifiers, preservatives, food dyes, added sugar, and removed fiber all produce the same health responses. Science showing correlations between UPF consumption and health outcomes also don't tend to show a dose response, which is odd. We'd expect lots of UPF consumption to be very bad and for some UPF consumption to be kind of bad, but we don't tend to see this in the data.

Nutritional research is also enormously difficult to perform. Any sort of controlled study is necessarily over short periods of time. Long term studies come with all of the messy confounds that make it remarkably difficult to determine causation.