[Matumio]

All general statements are false.

Joking aside, you can sometimes prove causality. It's just not as easy as proving a correlation. The most broadly accepted method would be a RCT (randomized controlled trial), though it is often unethical or unfeasible.

You can also build on causal assumption that everyone agrees on. (Like: A person's gender cannot be caused by a government policy. A person winning the lottery is not caused by anything other than playing the lottery.) From such knowledge you can build a causal graph, and (in some cases) draw new causal conclusions from statistical data.

Long story short: you cannot just dismiss a correlation as being useless for any proof of causality.