> 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.