I mean it is pretty odd that you would do a call for papers and pick out those you liked the most without knowing race or gender but then replace some just because it turns out they had the wrong gender/skin color/whatever.It's actually totally understandable.
Let me explain.
The conference organizers clearly have diversity as a goal.
In their naivete, they thought to themselves "we want equality for everyone! I know... let's just make this a pure meritocracy! Then there won't be bias and everyone will be happy!", and so they used a blind review process.
Of course, it didn't dawn on them at the time that if you took a blind random sampling of programmers, had them write presentations, and then approved them, you'd end up with a lack of diversity because the industry itself has an almost comically absurd lack of diversity.
So unless the average skill level of individuals in minority groups is extemely disproportionately higher than the average population (which rather contradicts the idea that race or gender don't impact skill), they would be poorly represented because they're poorly represented in our community in general.
So after approving their blindly reviewed set of presentations, they realized, shoot, they have an alarming lack of diversity in their presentations (read the above two paragraphs again if you don't understand why).
And so now they have to go back to the drawing board. All because they didn't understand population statistics.