The talk about pixels is misleading. The research paper doesn't mention pixels. They attached a 3x3 array of piezoelectric elements (roughly 3 cm in diameter each) behind a 13" OLED panel (one picture also shows a 5x4 array), using a sturdy frame structure to minimize interference of acoustic vibrations between the 9 (or 20) elements of the array.

Not to say that this isn't interesting, but it’s not display pixels emitting sound.

Right.

It's a way of putting speakers behind a display, which will probably be useful. This may improve bass response in laptops, because the speaker area is larger.