The null hypothesis is to assume there is innate equivalence; the burden of proof is on you to demonstrate otherwise and the attempts I've seen so far are lacking.
If there are very few women in computer science, you wouldn't assume it was due to some conspiracy or "systematic discrimination" without any proof. You wouldn't enact policies of reverse discrimination (affirmative action) to correct some supposed imbalance of the system.
You'd simply admit you don't know why, and search for proof of the reason.
Both "P" and "not P" can be null hypotheses, depending on the design of experiment.
The fact that we are all equal does not mean we are all equivalent. I suspect you may not be making this distinction.