You are misreading how Broader Impacts (BI) works. From your link:
> Some examples that illustrate contributions in each of the five areas are given below. Proposals need not address all of these areas, and PIs are advised to focus on those areas in which they are well prepared to make meaningful contributions.
"Broadening participation of underrepresented groups" is only one of the five areas, and no proposal was required to use it. I had proposals funded that focused on workforce development, for example. I saw others focus on science communication to the public (now forbidden in the memo this post is about!).
Proposals that passed grant panels were first and foremost always those that would great science. At ~10:1 oversubscription rates or more, proposals don't pass without it. The BI component needed to be credible but could be handled lots of ways.
Fundamentally, Congress recognized when defining BI as a component for merit review in the NSF that fundamental science only pays off in the long term. BI is a pragmatic choice to ensure that grants also yield near-term benefits to society as well.
>is only one of the five areas, and no proposal was required to use it.
Irrespective of whatever was going on in academia I take issue with this. Everyone who has a) a double digit number of brain cells b) has ever dealt with government approval in any capacity c) is't just a straight up liar knows that if the requirements set forth by a panel with discretionary authority says to do items 1-5 that you will not be approved without doing all of them, (unless of course you have the right last name or connections).
If you don't believe me watch any local board's meetings for the next 6mo and research everyone who comes before it after finding what outcome they got.
This has nothing to do with academia, DEI or what the other items on the list of requirements were. This is just how the sausage is made. It's all the same steps even if some factories are a little dirtier than others. So yeah, I 100% believe that if someone unconnected didn't pay the right lip service to the right things in every single one of the items in the list they would not get the outcome they wanted even if theoretically their stuff could have been approved with only 4/5 boxes checked. The approvers are not going to stick their necks out like that with no reason.