Mainly curious if they knew or thought it was likely that Allied people would be killed, and had they known for certain would that have changed their decision.
The next question is how often was friendly fire part of collateral damage. Was it unusual for strategic bombing to kill POWs? In general how many friendly fire incidents were caused not miscommunication or misfires, but known costs taken to damage the enemy.
They knew that POWs were likely in many major Japanese cities, but did try to avoid targeting cities that they knew had POW camps. In late July, they determined, from reports of Japanese POWs, that Hiroshima was the only city among the four final targets (Hiroshima, Kokura, Niigata, and Nagasaki) that was reported to not have a POW camp (which was incorrect). The Japanese POWs said that there was a camp in Nagasaki, but this could not be confirmed by photographs. The response to this was to confirm that Hiroshima would be the primary target; they also thought about different secondary targets but never went with them. They also said that any evidence about the camp locations should be used in determining bomb placement in the cities.
So, in short, they did take this into consideration, though it was not judged a big-enough issue to cancel any of the missions. It was not something they desired, but if you are aiming to destroy entire cities, the amount of collateral damage is necessarily going to be large, and they were well aware that it was possible that POW camps were in or near said cities.
This is an interesting topic, which I had never considered. An additional question on this topic. Do we know of any communication between the US government and the families of these POWs after the fact?