I mean, think of it this way: it comes down to how often you come across words in any language including English (even in ENG: you may forget how to spell words correctly if you don’t use or encounter them often), kind of the same logic with Kanji: a Japanese person doesn’t know all Kanji in the same way English speakers doesn’t know every single word that exists in ENG.
There are over 5000 Kanji but only about half of that is used in Japanese or closer to 2136 while the remainder consist of ones only present within technical jargon (medicine, science, politics, etc.). or certain Kanji only has limited uses in some words (but mostly written in kana). That is also accounting for grammar being “straight forward” more than English or Euro languages.
The “real” hard part is numerous readings (depending whether it’s paired with kana or another kanji, reflected from kunyomi & onyomi plus nanori when applied in people’s names). What I hate about most online translators is that it often gets lost in translation (like words used in the wrong context but on their own it’s correct, however not right for the situation or topic at hand).


it’s difficult for us Westerners, because the concept is weird to us at first, but when you get hang of it it’s not that hard, it’s a similar concept to how we don’t read words letter by letter but we just see them and know the meaning
Imo numbers are the best comparison, because just like with kanji, a number has multiple readings and a meaning.
A “2” can be read as “two”, “twe” or “seco” depending on how it’s used (2 vs 12/20 vs 2nd). Just off the top of my head.