The ventriloquism effect is what makes us think the ventriloquist's dummy is talking, even though our ears should be able to hear that they aren't. We can train ourselves out of it using a simple trick (but then we'd have to find out which dummies really are talking).
Although it's "the ventriloquism effect," it could easily be called "the movie effect," or "the concert effect," or "the effect that you get when you put speakers all around your living room but keep the tv front and center." When we hear a sound, and see motion that seems to correspond to that sound, we "hear" the sound as coming from the source of the motion. Our brains pull this on us all the time. It's what keeps babies fascinated with sock puppets and it's what keeps us from thinking about the fact that speakers all around a theater are meant to project voices that are supposed to be coming only from the front of the theater. Most notoriously it affects ventriloquists, and makes their lifeless pieces of plastic seem like they're actually talking.
Unless those "lifeless piece of plastic" are actually talking. But enough of that for now!
The effect is so common it goes unnoticed most of the time. For some time it was chalked up to humans naturally taking our primary cues from sight, and sight being our primary sense. In fact, we rely on the sense we are most confident about, which most of the time happens to be sight. To derail the ventriloquism effect, all we need to do is foster either confidence in our sense of hearing or doubt in our sense of sight.
We can tweak our sense of hearing with nothing more than a little training. Researchers showed participants in an experiment simultaneous flashes of light and bursts of sound. At first people tracked the light and thought the sound came from wherever the light appeared. Then the researchers started staggering the light and sound. They started out with a few seconds between the light and sound, which allowed people to easily distinguish between the location of the light and the location of the sound. The time between the light and sound was slowly whittled down, but people kept correctly tracking the noise.
And then there were the researchers who decided to take out their subjects' confidence in their own sight. This time subjects were shown circles on a screen accompanied by bursts of sound. People chose sight over hearing when it came to determining the location of the sound source. Then the circles started getting blurry, eventually becoming indistinct blobs. People started following their ears instead of their eyes, and located the source of sound easily.
So if you come up against a questionable dummy, you need do nothing more than squint your eyes and listen, or try to train yourself to see the discrepancy between the dummy and the ventriloquist. Either you'll be able to shake the illusion and see the ventriloquist is the one that's really talking, or you'll be able to shake the illusion and see that you're in a room with a demon dummy. At which point I suggest you run like hell and hope it doesn't crawl after you through the vents like a rat monster. Good luck!