Probably because it's totally asinine to say to someone "Hey you're black, come work for us!"
If you're suggesting that African American Computer Science majors should all get hired immediately out of school so that Google can increase its count from 1.0% to 1.02%, it doesn't work like that.
Two things in this discussion actually matter if you want to affect meaningful change:
1. The racial makeup of graduating Computer Science majors compared to the makeup of the US population. If only 2% of graduating CS majors are from a race that makes up 13% of the US population, you will never have anything approaching proportional representation.
2. The racial makeup of applicants compared to the makeup of those applicants extended offers. This is really the step where, if there are company-level systemic issues around race, they will show up. If 10% of your applicants are a given race but 1% are extended offers there might actually be an issue. This is much harder to determine because not only is it very hard to get the data, but you need a sufficient number of applications to get meaningful results. If you only get 100 resumes a year for a dozen positions, you won't get anything useful here.
Why not do it?