First is that copyright applies to written content (even short content such as tweets) and does not apply to factual data.
Second is that whatever rights may apply to the data, in the fintech scenario the users are scraping their own data.
So all kinds of copyright-specific laws, of which there are many, don't really apply in this case - and those are the laws that can easily get used against the service provider, unlike the possible ToS violation where the bank would have to against their own customers to enforce it.
That's the trick, they just put all the liability on the user for security issues. Most banks say you must never give your credentials to a third party
So these startups and fin-tech companies just slap a few clauses in their T&Cs to say it's the users problem, not theirs.
Banks benefit from certain bugs/holes in their software (e.g. shitty UI for viewing transaction history/balance -> more overdrafts -> more fees).
With copyright, why create something new when you can just sue somebody?