<bashes head repeatedly against wall>
The person sounding dense here is you!
It's nothing whatever, at all, even a little bit, to do with whether I, personally, like or dislike a process. The problem here is that the banks are training customers to do something fundamentally insecure.
A person is sitting minding their own business, when the phone rings and a strange voice says, "This is your bank. Please give me the following data used to establish your identity to banks."
According to the bank's thinking, the recipient in this situation should now give their identity data to the unknown incoming caller – to do with what they will.
Yes or no?
Obviously for the customer's security, the answer is No. Therefore the bank shouldn't expect them to. Shouldn't tell them to. Shouldn't have to manage not to get arsey because... they shouldn't be asking for the information in the first place.
If the security procedure the bank would prefer is that the recipient should put the phone down (and make a test call to ensure the line really has disconnected, because that's another scam) and call in to the published bank number THEN THAT IS WHAT THE INCOMING CALLER SHOULD INSTRUCT IN THE FIRST PLACE. Not encourage the person they're calling to do something fundamentally insecure, but then "permit" them to act more securely by calling back.
This is what I mean by banks' refusal to conceptualise that security is two-way.
Or, you know, I could just believe the banks are deliberately doing something they know full well is insecure for the customer but is more convenient for the bank, if you prefer? That they understand full well, but plan that the customer will bear the loss come the day that call is from a scammer, therefore from the bank's point of view this isn't a problem?