You see a third trace coming out of that plane, ground based on the silksreen, and leading to the two trace coming from the chip. So I would say, that two pins were grounded at one point, and they drill that grounding out. Maybe that was easier than making test points?
I think it is possible, they were mass programming them, like programming a whole panel at the same time, and then do the depanelisation. Maybe this was setting an address, or just pulling down the "ground this pin to program".