They get treated as duplicates because their content matches, not because their titles look alike. Google decides duplication by comparing the substance of the pages, the body text, not the title in isolation. So a similar title is at most a hint that prompts a closer look. If the pages underneath genuinely say different things, Google reads them as two distinct pages, and the matching titles do not trigger a duplicate verdict.
The flag actually fires when similar titles sit on top of similar content. The classic case is two thin location pages built from the same template with a city name swapped, where the whole page, not just the title, is near-identical. That is when Google folds them together, picks one version to index, and leaves the other out, because showing two copies of the same value would not help anyone. The shared title there is a symptom of the duplication, not its cause; the cause is in the bodies.
So the fix for a duplication worry starts below the title. Before you touch the titles, check whether the two pages’ content actually overlaps. If the bodies are distinct, the title resemblance is not your duplication problem, and rewriting the titles will not resolve a duplication that does not exist.
It is worth naming the real cost of matching titles even when the content is unique, because it is easy to conflate with duplication. Identical titles on genuinely different pages do not earn a duplicate flag, but they do weaken click-through, two identical-looking listings split the click and neither stands out, and Google may rewrite one of them to tell the pages apart. There is also a mild risk of the pages blurring on the same query. So vary the titles for the sake of clicks and clarity, just do not mistake that for fixing a duplication, which, if it is real, lives in the content rather than the title.