I'm in a pretty good mood right now, so I'll take the question as it is. I'm thinking that ACS is referencing Dawkins' latest book, where Dawkins says that Science can weigh in on the "god" question. ACS' statement seems to be a roundabout way of expressing Dawkins' point; at the same time, his statement is one that that gets put forth by theists as an 'atheist' belief (which always tends to rile a few feathers here).
It is possible to prove a negative, if you show that certain things that must accompany something else are not present; then it is shown: that particular something does not exist. As long as we're careful with our definitions, extremely careful that.
"There is no (normal) elephant in this room." Very easy. Normal (breathing) elephants, produced by currently understood biological means, leave obvious clues to their presence. We can see them. We can weigh the room I'm in on a scale, and then weight an completely identical room on a different scale, and compare weights. We can look for crushed furniture, smell the air, and check the bowl of peanuts to see if any are missing. All point to: no elephant in the room.,
The trouble for 'there is no god' is right off the bat, 'god' is an un-definable entity. All we can say is "if (un-definable entity) exist, with characteristic X, then Y should be present". All that leaves us with is an endless series of X's, with Y's to check for them. Since there are an infinite number of (un-definable entities), the whole exercise begins to get a little silly.
"How can you say that (un-definable entity) doesn't exist?" is thus a nonsensical question from the POV of inductive logic. Pragmatically speaking, it's not there until evidence begins to suggest it does. And the evidence would only speak towards the existence of definable entities, never undefined ones.