you are correct that it would still be interlaced when it goes to air . . but de-interlacing still creates a more “film” sort of look.
Interlaced footage from your camera has a different field (or half frame) 50 times a second (for PAL anyways). This 50 fields a second gives a very “video” look.
When you de-interlace in the most primitive of ways you are basically removing every second field or half frame. Then duplicate the one that is left so that you get a full frame. You’ve now got 25 frames every second, but you’ve lost vertical resolution. So when this goes to air it is still getting broadcast with fields but every two fields are exactly the same.
So instead of capturing a moment 50 times a second you’re now seeing it 25 times a second.
Instead of just duplicating each field there are options to re-create that missing field. AE does a pretty good job on it’s own, but plug-ins like Magic Bullet really are the best option.
Does that make more sense?