This is the Audio Illusion known as the McGurk Effect. Pay close attention to the video footage and watch what happens!

This effect was discovered completely by accident when Harry McGurk had his lab assistant dub over some video while studying how infants understand language. Because vision is our primary sense we actual perceive audio using our eyes first. Hearing, Taste, Smell and touch are all secondary senses when compared to sight.

Our brains take in the information we see and try to make the best guess as to what we're hearing as well. So in a nutshell, when we hear one thing but see another our sight takes precedence and we intrepret a different word being said.