There is a theory (theory! very theoretical) that the risk from radiation does not scale linearly and that low doses of radiation might actually be protective. See https://en.m.wikipedia.org/wiki/Radiation_hormesis
The “small dose radiation is actually beneficial” part is pure conjecture, and almost certainly wrong. But unrelated to the issue of the validity of the linear model.
Compare with the CTE, where a few bone-crushing traumas do not seem to cause it, but chronic long-term low grade impacts is more likely to cause it.
Another example, I probably could eat a massive amount of processed meat a few times without much ill effect, but years of eating moderate amounts might do you in.
J. Frank Parnell (from the movie Repo Man).
>In this multicenter trial, researchers enrolled 114 patients with moderate-to-mild knee osteoarthritis across three academic centers in Korea. Participants were randomly assigned to receive one of two radiation regimens — a very low dose (0.3 Gy) or a low dose (3 Gy)
This is gamma. Those doses are "low" because delivery here is localized. If given full-body the 3Gy has something like 25-50% mortality.
The cancer patient are delivered like 20-80Gy into the tumor and surrounding tissues which just kills cells outright.
The point is that "radiation" is multiparametric.