Is the connection between smoking and cancer a correlation or causality?
Depends on how you look at it.
If you try to prove "smoking->cancer" from the number of smokers that get cancer, that correlation is not very firm.
If you find that "90% of lung cancer patient have been smoking" you are talking about a very strong correlation. but that means turning the question around (it's called casuistic).
Even then you may have overlooked co-inciding variables like social class, age (!), etc.
Even though we all know that smoking isn't particularly healthy, I would call this a (strong) correlation with a strong hint to causality.