The point is that it's showing the reality of vision, not racism. There's a little black gun in that study that's very hard to distinguish in the hand of a sufficiently dark-skinned person. You're either going to take longer to decide (and the game has a timer) or you're going to make more errors.
There is no corresponding hard-to-identify threat in the hands of any of the white people.
Unless the white guy is wearing gloves, which evens the playing field, sort of.
Even in split second decisions, there is a decision making process. So, it must go something like this:
1. Can't tell if black person has gun because hands are same color as gun.
2. Shoot black person.
For this process to occur, there has to be an acceptance that guns are less visible in a black person's hand, as well as a black person is likely to have a gun, and likely to use it, but not include foregone conclusions about people of a certain race.
I am a little confused about why this is not about race. As I said above, white men can wear gloves. What if a white man has his hands in his jacket pockets. If uncertainty about whether a gun is present is the overriding factor, why is "there no corresponding hard-to-identify threat in the hands of any of the white people?" when it's very likely a white man can hold a hard to identify gun?