Most kanji are a combination of several smaller parts called "radicals" in English. If you look at these two kanji through this lens, you will see that it's actually a very simple mistake, one existing radical is replaced by another existing radical. It is very easy to imagine software that was working exactly like that: interpreting kanji as a combination of radicals rather than individual unrelated symbols