The job that the security org is doing for the board/owners/etc is controlling risk because they don't trust if their IT org says "everything is totally fine", and consider it likely that if the decision about whether to implement some activity or not is left purely to the IT org, then the convenience of implementation will override the level of risk control that the board/owners desire, so they choose to put these controls in a separate organization, empowering someone to say "Not until you do all this unwelcome work to implement it to the standards the company has chosen", preventing all the many mid-managers to save their effort or their costs by cutting corners at the expense of risks to "someone else's" money.
Of course, having these functions together is much more efficient in many ways! But the principal-agent problem is very real, especially so in large organizations, so that's why these choices get made this way, designing organizational structures that act as checks and balances on each other, not expecting everyone to magically cooperate for the greater good.